How does a society’s decision machinery begin to fail long before most citizens notice the damage? I introduce a dynamic, attribute–driven network model in which every node carries two variables—individual fitness F and institutional coverage C. Coverage is the probability that an agent’s data are ingested by optimisation pipelines (markets, recommender systems, policy feedback loops). A simple feedback rule links the two: as inequality widens, the expected coverage of low-fitness nodes decays faster than that of high-fitness nodes, causing the data stream that powers collective decisions to shrink and skew. The resulting cascade concentrates information, attention, and algorithmic leverage inside an ever-smaller elite core—a phenomenon I call the technology gravity well.

Analytical bounds identify a tipping region beyond which coverage collapses abruptly; below that region, decay is slow enough for corrective intervention. A Monte-Carlo agent-based simulation (1 000 nodes, preferential-attachment baseline) demonstrates three regimes: (i) stable coverage when disparity is modest, (ii) gravity-well implosion once the tipping region is crossed, and (iii) stabilisation when a decentralised functional-model intervention exposes the input–state–output mapping that drives each optimisation pipeline. No wealth transfers or exogenous rewiring are assumed; the intervention acts by restoring visibility, not by redistributing assets.

Because coverage is treated as an endogenous network variable, the model links micro-level attribute dynamics to macro patterns usually discussed under segregation, polarisation, or epistemic drift. Crucially, it predicts that any society running opaque optimisation logic is pulled toward widening disparity and systemic information loss, making the gravity well a general failure mode that spans economic management, climate policy, biosecurity, and other existential-risk domains. Reversing the collapse therefore requires treating the functional model of intelligence itself as a shared public good—transparent, auditable, and open to recursive self-correction.

The CEPII BACI database (link) offers a comprehensive dataset on bilateral trade flows between 200 countries, covering 5,000 products from 1995 to 2023. This dataset serves as a "gold mine" for constructing and analyzing weighted trade networks. We will use selected networks from BACI to demonstrate various exploratory analysis techniques for weighted networks.

The main challenges in weighted network visualization and analysis are:

• Large, dense networks (n > 20) are difficult to visualize clearly using traditional graph layouts. A more effective approach is a matrix representation with a "right" node ordering.
• A large range of weight values and a highly skewed weight distribution can distort analysis. Applying monotonic transformations (e.g., log scaling) helps adjust weights while preserving the importance of links.
• Dominant nodes (e.g., the U.S., China, Germany, the UK) often overshadow smaller economies. To ensure comparability, we can:
  • Normalize networks (using Markov or Balassa transformations)
  • Use size-independent measures (e.g., the Salton index)

To uncover the underlying structure of large or dense networks, we can reduce them to their skeleton by removing less significant links or nodes. We examine three types of skeletons:

• Closest k-neighbors skeleton (retaining only the strongest links)
• Pathfinder skeleton (preserving critical shortest paths)
• Weighted degree cores (focusing on nodes with high interaction intensity)

A time series of trade networks (e.g., annual snapshots) on a selected topic forms a temporal network. By inspecting the evolution of network skeletons, we can identify structural shifts in global trade patterns over time.

We analyze longitudinal weighted network data by proposing an extension of the stochastic block models designed to handle count data characterized by an excess of zeros. Specifically, we explicitly model the distribution of the dyad referred to each pair of nodes, conditional on the blocks they occupied at each time occasion. The time varying block memberships are assumed to evolve over time according to an unobservable Markov chain. For the conditional distribution of the dyad, we consider two alternative versions of the bivariate Poisson distribution: a classical zero-inflated model, and a hurdle model formulation. We further extend the framework to multilayer networks, in which the same set of nodes can interact through different types of connections, organized into distinct layers.
For the resulting dynamic models we introduce variational inference as an approximation of likelihood inference. The overall approach is illustrated by two applications: one based on patient transfers within a network of Italian hospitals, and another using data from the SIPRI Arms Transfers Database referred to international arms transfers.

The messaging service Telegram has become popular among criminals as a learning environment and online marketplace. Massive group chats and end-to-end encrypted secret chats enable criminals to exchange information and sell illegal goods, among which firearms, to a large audience. This study aims to investigate the process of illegal firearms trading on Telegram through the application of crime scripting and social network analysis. The Telegram data extracted from one specific smartphone belonging to one of the leaders of a criminal organisation (N = 14) involved in illicit firearms trafficking was used, which was seized during a Dutch police investigation. The user of this phone sent and received 126,589 chat messages on Telegram in total of which 119,230 messages in group chats (N = 17) and 7,359 messages in private chats (with 90 unique accounts) over the span of 377 days. This study will address several research questions. First, the study will determine the user’s level of activity across the 17 group chats. The content of both group chats and private chats will be analysed to find out what topics were being discussed. Furthermore, the study will examine whether firearm trade deals were exclusively made in private chats or also within group chats. Additionally, the extent to which private chats with possible buyers or sellers arose after the user posted or responded to firearms advertisements in group chats will be explored. Lastly, the study will identify the different scenes and facets involved in the process of firearms trafficking on Telegram.

Social Representation Theory (SRT) provides a robust framework for investigating how social groups construct shared systems of meaning, particularly through discourse. At its core lie two fundamental cognitive and communicative processes: anchoring, whereby new information is assimilated by linking it to familiar categories, thus enabling interpretation and classification; and objectification, through which abstract concepts are shaped into concrete and perceptible entities, facilitating their incorporation into everyday thought. As discourse increasingly unfolds in digital environments, where narratives are rapidly generated, transformed, and contested, understanding how social representations emerge and evolve poses significant theoretical and methodological challenges.
To address these challenges, we introduce ThemeScope, an analytic strategy designed to detect and visualise social representations in large-scale digital corpora. Bridging qualitative theory and computational techniques, ThemeScope combines co-word analysis, network-based community detection, and visualisation tools to uncover thematic structures in digital discourse. The framework introduces two novel quantitative measures: the Prototypical Salience Index, which captures the relevance and embeddedness of a concept within discourse (anchoring), and the Concreteness Score, which quantifies the extent to which abstract topics are articulated in tangible forms (objectification). These metrics position topics on a two-dimensional space defined by structural salience and semantic concreteness, aiding the identification of stable cores, ideological constructs, emerging practices, and latent representations.
We illustrate the potential of this approach through a case study on the Israel-Palestine conflict, analysing social media discourse between October and December 2023. The results highlight the coexistence of multiple representational fields shaped by competing narratives of identity, morality, and legitimacy.
ThemeScope constitutes a theoretically grounded and computationally scalable framework that allows the systematic investigation of collective meaning-making processes across large and dynamic digital environments.

Social media are a fundamental arena in the context of electoral campaigns, offering a new dynamic space for political communication and public engagement. During elections, these platforms are not only used by political actors to share and disseminate content, but they are also a space where users interact, express their opinions, and influence each other. User comments, in particular, represent a valuable source of information for analyzing patterns of interaction and the circulation of sentiment within the online public sphere.

In this study, we investigate user interaction in two different social media platforms in the two months preceding the 2024 European Election in Italy, covering the period from April 2024 to June 2024. We retrieved the posts and the comments of different political leaders. Based on this data, we extracted the comments and constructed a direct one-mode network among users that captures the structure and the intensity of comment-based exchanges.

To enrich the analysis, we applied sentiment analysis techniques to each comment, allowing us to assign a polarity score to interactions. This enabled the construction of a signed network, where edges represent either positive or negative sentiment between users. We then tested structural balance theories on the network of interactions. This provided a framework for examining not only the topology of online political discussion but also the affective dynamics that characterize user engagement during electoral periods, making it possible to compare the different dynamics present in each profile and identify processes of polarization among user interactions.

We present an analysis of the complex network structure of some microblogging platforms, comparing their internal organizations and, for the case of BlueSky, its growth during a period of massive migration from X/Twitter. Topological differences are the result of platforms’ functionalities and of individuals’ behaviors. At the same time, the sudden increase of users in BlueSky acts as a large perturbation to the platform’s dynamics and provides a unique opportunity to explore how the mesoscale organization in a decentralized environment changes, giving rise to scaling laws. Using a dataset of 21 million users, we investigate the differences introduced by the migration “shock” and compare BlueSky’s network features with those of other platforms, including Truth Social, Gab, Mastodon, Parler, and X/Twitter, as well as a previous snapshot of BlueSky updated to March 2024. Notably, BlueSky’s use of “starter packs” facilitates connectivity and may significantly impact user behavior and collective behavior. Preliminary results reveal the emergence of scaling relations between the number of followers and followees, such as which is of interest due to its effect on the bursty dynamics of online collective attention[1]. Additionally, we report on higher-order topological correlations, conveying information on the hierarchical structure[2], and (dis)similarities in mesoscale organization, such as community structure and network robustness, specifically comparing Truth Social with BlueSky. This ongoing study contributes to understanding how users behavior impacts the emergence of network topology and how the decentralized microblogging platforms adapt to external, large-scale, shocks and evolve under non-equilibrium conditions, offering insights into the interplay between structure and dynamics in complex social networks.structure of the mobility network varies depending on the approach used to model the spreading process.

[1] De Domenico, Altmann Scientific Reports 10, 4629
(2020).
[2] Artime, d’Andrea, Gallotti, Sacco, De Domenico, Scientific Reports 10, 14392 (2020).100 101 102 103

Urban landscapes around the world feature the coexistence of stable drug markets and violence. However, the correlation between volumes of drug dealing and violence is frequently inconsistent across areas and time. In this work, we hypothesize that such inconsistency is due to the spurious nature of the relationship, and we pose that an important missing link in the transmission chain is given by organized crime groups (OCGs) dynamics. Our starting point is that when an illicit market is contendible, with multiple groups potentially competing in the same area, OCGs may resort to cooperation as a mechanism to mitigate the risks of unbounded competition. Therefore, when cooperation falls apart, business opportunities (i.e. drug dealing) will be eroded and violence will raise. We test our theory on a novel dataset made of the complete crime dataset of OCG-related offenses of Merseyside, U.K., integrated with neighbourhood-level socio-economic data. We show that, net of urban and socio-demographic factors, a 1% increase in inter-OCG cooperation is simultaneously and strongly associated with a 1.4% future increase in drug dealing and a 1.5% decrease in violence at a monthly level for each neighborhood. We complement this result with sharp prediction on the emerging structure of the OCG cooperation alliances.

Collaboration is a fundamental feature of crime that is defined as organized. Despite considerable research focusing on collaboration within organized crime groups in recent years, only a few studies have examined cooperation between members of different groups or clans of large criminal organizations. Collaboration facilitates coordination and resource acquisition, yet it is challenged by the inherent lack of trust characterizing the criminal environment. This study contributes to the literature by investigating dynamics of collaboration among members of different ‘Ndrangheta clans, known as locali. Specifically, it adopts a multiplex approach to examine the factors associated with collaboration both within and among locali, emphasizing the role of kinship compared to the one of leadership and shared membership, proxies for the formal organizational structure. Data were collected from three major Italian law enforcement investigations targeting multiple ‘Ndrangheta locali and analyzed using multiplexity-related descriptive statistics and quasi-Poisson multivariate quadratic assignment procedure regressions to test the hypotheses. The findings suggest that while the formal hierarchy is primarily associated with collaboration across locali, kinship fosters cooperation within locali, also showing some indirect association in inter-locali interactions.

The UN Agenda 2030 highlights how corruption is one of the factors hindering sustainable development. However, to prevent and combat corruption, it is essential to deeply understand this latent phenomenon, which is not easy to measure. A step forward has been made by the Italian Anti-Corruption Authority, which provides the public with 70 indicators capable of measuring the risk of corruption across territories. Specifically, the methodology involves also the organization into pillars and the use of composite indicators that simplify the interpretation of the complexity arising from multiple dimensions by reducing it to a single number. But some methodological approaches have also been experimentally developed that contemplate the use of sociale network analysis, with particular reference to public tenders. The presence of information on individuals who have assumed top positions in organizations allows us to address, among other things, the issue of conflict of interest, revolving doors, as well as identify the awards presumably characterized by a higher risk of corruption. In particular, one in-depth analysis focused on the creation of a database comprising public tenders issued by healthcare organizations, the companies awarded the contracts, and the individuals holding senior positions in both types of entities. Based on this data, a network was constructed to capture the full set of individuals and organizations involved in the issuance and awarding of public tenders within the Italian healthcare sector, along with the relationships among them. In light of the relationships thus identified, it becomes possible to flag a number of contract awards that are presumably associated with a higher risk of corruption. Finally, a general methodology has been developed that is functional for mapping the shareholdings, family ties and economic interests of individuals with public positions or other individuals subject to a declaration of absence of conflict of interest.

This paper reports results from a project to develop a classification system for criminal groups that is more systematic and empirically based than the rather ad hoc collection of categories that is currently in use. The population of criminal groups to be classified was constructed from data on all co-offences reported by police in Canada during 2006-2009. These co-offences induce a 1-mode criminal network consisting of approximately 630,000 alleged co-offenders connected by approximately 800,000 edges.
This network has approximately 163,000 components, or criminal groups, ranging in size from 2 to 198 persons, with one giant component of approximately 172,000 persons. This giant component was decomposed by Louvain community detection into 478 smaller groups ranging in size from 2 to 2,266 persons. Each of the approximately 164,000 criminal groups was characterized by 32 variables describing the group’s size, the volume, duration, jurisdictional reach, and types of its criminal activities, its members’ attributes, and the structure of co-offending ties among the members. Conventional hierarchical clustering identified 5 main clusters of groups, with two of the clusters each subdivided into 5 sub-clusters. Each cluster and sub-cluster was then characterized by mean scores of the constituent criminal groups on the classifying variables.
Work continues on refining the methods of decomposing the criminal network and clustering the resulting criminal groups.

Generative Artificial Intelligence (GAI) text production is crucial to research fields as Data Science (DS) and Network Textual Data Analysis (NTDA), the main purposes of GAI being to simulate human language production, exploiting both Machine Learning (ML) and Large Language Models (LLMs).

However, as pre-trained probabilistic models, LLMs are biased when built on non-perfectly balanced data as for retrieval sources, taxonomy, ontology interconnections and linguistic inference. This is most relevant to DS and NTDA, as it can contribute in social media to spreading fake news, conspiracy theories, counterproductive narratives, and online hate speech. Equally relevant is GAI being devoid of a reality formal model, causing GAI to have no ethics, as it cannot identify and correct its inaccuracies. This brings LLMs and GAI to suffer from effectiveness and reliability issues, showing tendency to prompt incorrect and discriminatory information, and hallucinations.

Newborn Neuro-Symbolic Artificial Intelligence (NSAI) tries to cope with these issues building elementary ontologies to integrate human symbolic reasoning principles with ML and Artificial Neural Networks (ANNs). Here we will demonstrate that better results come integrating also formalized morphosyntactic and semantic information, as those relating to Italian negation grammar. Therefore, to tackle on-line hate speech, we propose here a method of Sentiment Analysis (SA) that uses NooJ software to build formal ontologies and syntactic grammars within graphs representing finite state automata/transducers. While ontologies will conceptualize sets of word having contiguous contextualized meanings, syntactic grammars will parse texts using Italian formalized morphosyntax and semantics.

The rapid rise of social media platforms has created new opportunities to understand public sentiment on complex social issues like migration. Unlike traditional methods, such as structured questionnaires, that have long been used to uncover underlying attitudes, social media offers a rich and evolving source of unstructured text that reflects real-time, spontaneous public opinion.
Surveys that combine open-ended questions with structured scales, such as the Semantic Differential and the Bogardus Social Distance Scale, provide a way to quantify these underlying attitudes. When analyzed through models like the Graded Response Model (GRM) within the Item Response Theory (IRT) framework, these instruments can reveal traits about respondents' views on migration. However, the informal and organic nature of social media discourse offers an additional, valuable layer of insight into how people express and share their attitudes in public forums.
This study investigates the relationship between latent attitudes captured through traditional questionnaires fulfilled through social media platform and the semantic patterns present in open-ended responses of the same questionnaires. By constructing a semantic network from these responses, we assess how closely the structure of the network correlates with the latent traits estimated from questionnaire data. This evaluation allows us to determine whether the semantic network encodes meaningful information about respondents' attitudes. If so, it opens the door to semi-supervised learning approaches that can infer latent traits directly from textual data, offering a scalable alternative for attitude measurement beyond conventional survey methods even on social media content and comments.

Research on online discourse on marginalized communities has largely concentrated on hate speech, often overlooking the role of self-representation and community-driven narratives. While the damaging effects of hateful rhetoric have been well-documented, less attention has been paid to how targeted groups articulate their identities, assert cultural agency, and construct counter-narratives in digital spaces.
In this work, we focus on how marginalized communities, particularly Roma italian activists and support networks, actively engage in digital self-definition. Drawing on anthropological methods, this study explores the symbolic frameworks, narrative strategies, and themes that shape these auto-narratives. Anthropology, with its long tradition of examining cultural identities, coupled with quantitative analysis, provides essential tools for identifying recurring topics, metaphors, and symbolic patterns that characterise how communities contest exclusion and reclaim representation.
This work combines digital ethnography, discourse analysis, social and semantic network analysis to systematically examine the language and narrative structures used by Roma influencers and community platforms. In addition, network analysis is employed to investigate the structure and dynamics of digital support communities, revealing how relational ties contribute to the circulation and reinforcement of counter-discourses.
By integrating qualitative and quantitative approaches, the study aims to investigate how Roma digital communities resist stereotypical representations, negotiate identity, and contrast hate in online environments. Ultimately, it demonstrates the need for interdisciplinary approach to decode the cultural and rhetorical complexity of digital self-representation and to better understand the role of marginalized voices in shaping contemporary digital discourse.

A central task in network analysis is to model social influence, that is, how the social environment shapes individual behaviors and outcomes. Autologistic actor attribute models (ALAAMs) provide a relevant framework for this purpose, using an exponential-family formulation to model binary actor-level outcomes given an observed network. However, ALAAMs treat the network as a fixed, deterministic object, without modelling its structure.
Latent position models, by contrast, represent network data in a low-dimensional latent space, where actors positioned closer together are more likely to share a social tie. These models are flexible and can incorporate dyadic covariates, node attributes, and other structural effects.

In this work, we propose a novel approach that bridges these two modelling frameworks by jointly modelling the outcome as a function of network structure, represented through a latent social space. Specifically, we introduce the logistic actor-attribute latent space model for social influence.
Our goal is to model the probability of a binary actor-level outcome as a function of both observed covariates and latent positions. The latent space serves as an interpretable, low-dimensional representation of the underlying social structure. Conditional on these latent positions, outcomes are assumed to be independent, with the latent space capturing complex dependencies not explained by covariates alone.
The model is formulated within a Bayesian framework, with inference performed via a Gibbs sampling algorithm, enabling efficient posterior estimation and principled uncertainty quantification.

Recent advances in technology have made it easier to collect large and complex time-stamped relational hyper-event data. This type of data captures events involving more than two entities at the same time. Relational Hyper Event Models (RHEMs) aim to explain the dynamics of these events by modeling the event rate as a function of statistics based on history and external information.

However, despite the complexity of the data, most current RHEM approaches still rely on a linearity assumption to model this relationship. In this work, we address this limitation by introducing a more flexible model that allows the effects of statistics to vary non-linearly and over time. While time-varying and non-linear effects have been used in relational event modeling, we take this further by modeling joint time-varying non-linear effects using tensor product smooths.

We validate our methodology on both synthetic and empirical data. In particular, we use RHEMs to study how patterns of scientific collaboration and impact evolve over time. Our approach provides deeper insights into the dynamic factors driving relational hyper-events, allowing us to evaluate potential non-monotonic patterns that cannot be identified using linear models.

Mathematical research in Russia has had a substantial impact on multiple areas within mathematics as well as on disciplines beyond it. The work explores the publication patterns of Russian mathematicians using data from the Web of Science (WoS) database, aiming to understand how their contributions have evolved over time and what factors have influenced these dynamics.

We analyze publication trends across various subfields (both theoretical and applied), highlight the shifts in international collaboration, and examine citation impact as a proxy for global influence. Special attention is given to periods of decline and resurgence, and particularly the post-Soviet era and the recent renaissance in foundational areas such as topology, dynamical systems, and algebraic geometry. The study considers how geopolitical shifts, institutional changes, and evolving research priorities have shaped scientific output. Moreover, we apply new centrality measures that take into account parameters of vertices.

The research also seeks to identify leading figures and institutions, offering insights into how Russian mathematics maintains its relevance in the global research landscape. We aim to better understand not only the trajectory of a national scientific tradition but also the broader, non-linear nature of mathematical development.

This study demonstrates the pedagogical value of Systematic Literature Network Analysis (SLNA) for teaching Social Network Analysis (SNA) across disciplinary boundaries. By applying SLNA to the field of in-work poverty research in Europe, the study exemplifies how computational network techniques can be integrated into higher education curricula to foster critical engagement with academic knowledge structures. SLNA combines bibliometric data with citation, co-occurrence, and co-citation network analysis, enabling students to explore the intellectual landscape of a research field and to identify influential works, thematic clusters, and emerging areas of inquiry.
The analysis draws on 822 Scopus-indexed publications, filtered through methodological and geographic criteria relevant to European labor and social policy. Tools such as VOSviewer, Pajek, and Biblioshiny are employed to construct and interpret various networks, perform main path and burst detection analyses, and map the thematic evolution of the field. In the context of generative AI, SLNA offers a structured, transparent, and reproducible framework that complements algorithmic summarization tools while preserving analytical depth and interpretive agency. Additionally, AI aids in collecting necessary keywords that are at the foundation of a literature review search process.
The study focuses on in-work poverty, a complex and policy-relevant phenomenon that exemplifies the intersection of employment, social inequality, and welfare systems. The results reveal a dynamic research trajectory moving from foundational discussions of poverty measurement to contemporary debates on precarity, minimum wages, and gendered vulnerabilities. This pedagogical application of SLNA supports interdisciplinary learning outcomes by teaching students to navigate and interpret large-scale academic corpora through network logic. The study reflects on the didactic opportunities and limitations of teaching SNA through SLNA, including software accessibility, data literacy requirements, and the potential for reflective, inquiry-based learning.

Charitable food provision (CFP) has grown into a structured yet contested component of urban welfare, increasingly reliant on cross-sectoral partnerships and surplus redistribution mechanisms. This contribution presents a mixed-method protocol for reconstructing the emergence and consolidation of CFP fields over the past 25 years in Italy, the Netherlands, and Japan—three countries marked by contrasting institutional regimes but convergent trends in welfare hybridization and the discursive bridging between food insecurity and the ecological sustainability transition.
Anchored in the combination of Strategic Action Field (SAF) theory with social network analysis, the paper conceptualizes CFP as an evolving organizational field shaped by competition, cooperation, and nested dependencies between food charities, state institutions, corporate actors, other civil society organizations, and the urban poor. We adopt event sequence analysis (ESA) to unfold field dynamics using an original dataset built from multi-source textual data (e.g. news archives, legal records, organizational reports) gathered in each country. We consider a wide set of events (e.g. organizational founding, partnership creation, regulatory changes, advocacy campaigns, policy implementation, episodes of contention, etc) which are linked through intentional dependencies and/or shared common ground.
Our ESA data collection protocol is designed for high interoperability with other longitudinal network-analytic approaches, enabling the construction of time-stamped bipartite and one-mode networks (e.g. event-to-actor, event-to-event, actor-to-actor). Building blocks of field formation are identified via modularity analysis, while blockmodeling techniques are applied to map actor roles and country-specific pathways. The presentation will showcase early-stage insights from the Italian case and discuss the protocol’s potential for comparative research and field-level policy evaluation.
By bridging event analysis with network modeling, the article contributes to methodological innovation in the study of community welfare ecosystems, offering tools to trace, compare, and explain the emergence of complex support infrastructures across diverse urban contexts.

Community interventions addressing violence against women often reflect a fragmented governance model, where collaboration among specialized (anti-violence centres) and generalist services (social service, hospital, police) remains weak. This fragmentation hinders the development of integrated multi-agency responses, limiting the capacity to mobilize social and institutional resources in support of women. This paper presents findings from a nationwide research project exploring how inter-organisational networks between public and private actors influence the design and implementation of local anti-violence policies. The study focuses on three Italian regions and adopts a mixed-method approach that integrates Social Network Analysis (SNA) with qualitative interviews and document analysis.SNA is employed both descriptively and analytically. At the descriptive level, sociometric techniques (ego-whole networks) are used to map formal and informal relations among organizations operating in the field of gender-based violence. At the analytical level, we apply centrality measures to nodes and the network, as well as density and embeddedness indicators (transitivity and reciprocity), to assess coordination patterns and network cohesion.

Results confirm the tendency to favour modes of intervention based on aggregation rather than on integration of objectives according to an intersectional and systemic logic. The network structure shows high levels of dyadic collaboration and a weakness of multilateral ties, typical of territorially organised “systems”, in the form of partnerships or protocols between several subjects. As confirmed by levels of transitivity, there is a marked polarization of the network around the nodes directly involved in taking charge of women and implementing interventions in the field.

This contribution demonstrates how SNA can support the analysis and strategic development of collaborative ecosystems in social policy domain. By identifying patterns of cooperation and disconnection, SNA offers a powerful toolkit for enhancing territorial governance and generating social capital in complex intervention fields, such as those relating to anti-violence policies.

In the rural area of Capitanata (FG), agricultural labor exploitation and illegal intermediation systems persist due to weak welfare structures and limited institutional coordination. In response, an informal inter-organizational network of third sector and civil society actors has emerged to support and protect migrant workers’ rights.
In this context, this contribution aims to explore the role of inter-organizational networks active in Capitanata, emphasizing how interdependence among actors can act both as a constraint and a resource for collective action. The research originates from an experimental action-research thesis conducted in 2023, which investigated the structure and functioning of the network through a mixed-method approach combining Social Network Analysis (SNA), qualitative interviews, and a period of participant observation in Borgo Mezzanone, an informal settlement where migrants live in conditions of extreme marginalization.
The analysis carried out a fragmented collaborative ecosystem, with wide margins of improvement and synergistic potential. Informal relationships activated by local actors could form a basis for developing more cohesive and effective community interventions. Therefore, the strengthening of internal coordination is configured as a crucial strategy to avoid duplication of activities, optimize available resources, and promote more integrated action. The network would benefit from the introduction of planning mechanisms capable of orienting initiatives towards more strategic, structured, and sustainable cooperation in the long term.
Starting from this analysis, possible scenarios for strengthening the network are imagined, aimed at promoting more effective collaboration and greater protagonism of the actors in policy-making processes. From a future perspective, a desirable evolution of the study could consist of activating a participatory path between public and private bodies, aimed at the co-planning of urban and community regeneration interventions. That path, supported by a third party with technical expertise, would respond in a shared and inclusive manner to the needs of the territory.

Relational event models (REMs) are statistical models for the analysis of sequences of relational observations. Estimating these models is straightforward (albeit potentially computationally demanding) when the exact sequence of events is known, such as in the case of phone call records. However, sequences of relational event data may be incomplete or difficult to gather, e.g., when online platforms only provide aggregated data on user behaviors or when networks derive from cognitive perceptions of relationships, like friendship perceptions. Those perceptions are typically collected through survey panels that do not gather instantaneous network changes.

We present an application of the Expectation-Maximization algorithm to estimate the parameters of a REM variant, the Dynamic Network Actor Model (DyNAM), when the sequences are partially observed. We consider the evolution of an incomplete sequence as the result of two conditionally independent DyNAM processes, one governing the tie creation and the other the tie deletion. The generalized Expectation-Maximization is considered for imputing missing data and generating a sequence of proposals that do not necessarily maximize the objective function. The Expectation step is computationally intractable and approximated by random samples of plausible event sequences aggregated following Multiple Importance Sampling schemes.

Additionally, we show that this implementation is not limited to the analysis of incomplete sequences of relational events but can also be applied to analyze the co-evolution of network panel data and relational event data. We illustrate the method by analyzing data on friendship networks collected in a panel design and online interactions.

The E-I index, defined as the number of between-group (external) minus the number of within-group (internal) edges divided by the total number of edges in a network, is widely used as a measure of homophily in social network analysis. Smaller values of the E-I index indicate greater homophily. Permutation tests are commonly used to test hypotheses involving the E-I index. We will demonstrate that the results of these tests are often misinterpreted and present an analytical expression for the null distribution of the E-I index in random graphs with a given group distribution and a specified number of edges. Using the classical dataset on the information flow between organizations, collected by Knoke and Wood in 1978, we argue the importance of selecting the correct reference value for E-I index interpretation and how the choice of null distribution can impact the results of significance tests. We discuss how our results extend to E-I indices that measure homophily at the group and individual level.

This study seeks to investigate how two basic types of network ties – friendship and romance – relate to one another over time at the network level. It is hypothesized that specific taboo and eligibility norms influence the co-evolution of the two networks by systematically constraining which new friendship ties may form given a person’s past/existing romantic ties and which romantic ties may form given a person’s past/existing friendship ties. We further compare how the norm mechanism works differently when friendship interacts with romantic ties as contrasted to when friendship interacts with (non-romantic) sexual ties. Alternative explanations such as propinquity and homophily are also considered and tested. The study uses the Romantic Pairs Add Health Data for the longitudinal romantic/sexual ties network and the Restricted Add Health Data for the longitudinal friendship ties network. Data are analyzed using stochastic actor-oriented models (SAOM) where constraints are construed as conditional, multiplex propensities for tie-formation and deletion.

The mobility of modern academic scholars offers valuable insights into the dynamics of career development, institutional reputation, and the strategic choices shaping academic trajectories.
In this study, we examine the scholarly mobility of Italian academic statisticians between 2000 and 2024, with the aim of uncovering the drivers of the academic mobility process.

Using multiple data collection strategies and machine learning-based preprocessing, we built a comprehensive dataset that captures mobility patterns over time. The data includes names, surnames, and institutional affiliations from the Italian Ministry of University and Research (MUR). Subsequently, we identified each scholar’s Scopus ID and extracted their complete publication history from the past 24 years. To address missing affiliation data—particularly for scholars who moved abroad—we used institutional information available in Scopus records.

In addition to tracking individual mobility, we collected information on each scholar’s scientific disciplinary sector, the institutions involved, and their publication output. The resulting dynamic network was analyzed using the Relational Event Model (REM), which allowed us to investigate how individual characteristics, institutional factors, and research productivity shape mobility trajectories over time. A range of covariates was used to explore the drivers of network dynamics. We address key questions such as: Does co-authorship influence mobility, or does mobility shape research productivity? Do scholars tend to move to institutions where their co-authors are based?

In the age of hyperconnectivity, digital environments increasingly influence emotional states specifically social media platforms consuming algorithmic curation for maximizing user engagement. This paper presents the notion of the Digital Mood Swing a concept characterizing sudden, often unforeseen changes in emotional well-being because of digital interaction. Based on interdisciplinary insights from psychology, data science, and media studies, we discuss how content exposure, notification design, feedback loops (likes, shares, comments), and algorithmic biases trigger sudden emotional swings from elation to anxiety or despair. By means of qualitative interviews and quantitative user behaviour analysis, we identify patterns between screen time, content type, and platform features and mood volatility. For mental health, especially for vulnerable populations and adolescents, the paper also suggests ideas then offers plans to intervene and policies to recommend for digital well-being and lessened emotional manipulation. This work ultimately calls for a deeper awareness of the psychological architecture of digital platforms as well as the need for more ethically aligned design.

While previous research has emphasized psychological and behavioural impacts of digital consumption, this study uniquely integrates Social Network Analysis (SNA) to investigate how structural relationships within digital networks affect users’ emotional states. Through qualitative interviews and quantitative behavioural data, including relational metrics derived from SNA, we uncover patterns linking user engagement, network structure, and emotional response. This work ultimately calls for a deeper awareness of the psychological architecture of digital platforms as well as the need for more ethically aligned design. Our analysis focuses on how specific relational dynamics, including network centrality, density, and clustering, influence the emotional consequences of online activity. The study highlights heightened susceptibility among adolescents and vulnerable groups, raising ethical concerns about digital emotional manipulation.

Artificial Intelligence (AI) is now having a major influence now on the shaping of decisions within important areas such as jobs, education, healthcare, and public services. While it is often seen as a tool that brings speed, accuracy and neutrality. In recent research shows that AI can also carry forward and increase existing social inequalities. This happens mainly because AI systems learn from data that reflect real world unfairness such as discrimination based on caste, class, race, gender, or ability.

This paper reviews and discusses how AI technologies may unintentionally affect people from marginalised communities. It explains how bias enters AI systems through historical data, design choices and lack of proper testing. Case studies from different sectors like recruitment, law enforcement and education are used to highlight real life problems caused by AI tools. These examples show that AI systems can favour certain groups while putting others at a disadvantage, regularly without anyone noticing it.

In these paper explores the use of Social Network Analysis (SNA) as a method to better understand how AI influences social relationships and power dynamics in digital spaces. SNA helps reveal which communities gain visibility and influence through AI-driven platforms and which groups remain excluded or marginalised, offering a clearer picture of digital inequality beyond just biased algorithms.

The study argues that treating AI as just a technical tool is not enough. Instead, it should be seen as part of a larger social system, influenced by human values, power structures and historical context. To reduce the harm caused by biased AI, the paper suggests including human oversight, community participation and transparent decision-making in the development and use of AI. The review also points out gaps in current knowledge and offers directions for future research and policy changes.

The rapid convergence of quantum computing and artificial intelligence (AI) signals a foundational shift in the development of next-generation digital systems. Beyond technical advancement, this convergence holds critical implications for cybersecurity, particularly in quantum cryptography, post-quantum security, and algorithmic risk mitigation. Although academic output in this domain has grown rapidly, there remains limited understanding of how institutional actors are structuring this convergence and how their research aligns with sector-specific priorities, especially in cybersecurity.

This study applies Social Network Analysis methodologies to examine institutional relationships and knowledge diffusion patterns within the quantum-AI research landscape. A systematic review is conducted using the PRISMA framework on identified literature from reputed journals and conference proceedings. Employing network visualisation and transformer-based natural language processing models, the study constructs semantic co-occurrence and institutional affiliation networks to trace conceptual evolution and collaborative network structures. Key research themes such as variational circuits, quantum kernels, explainability, and optimisation are extracted and analysed to identify domains of cognitive convergence.

Institutional affiliation analysis classifies academic, governmental, and corporate contributors. Co-authorship patterns and topic relevance are examined to assess alignment with cybersecurity-related sectors. The resulting network structures highlight patterns of concentration, collaboration gaps, and emerging translational actors shaping the interface between research and applied cybersecurity innovation.

By revealing the institutional architecture and sectoral trajectories of quantum-AI research, this study contributes a network-informed framework for understanding how interdisciplinary innovation diffuses from research environments into security-focused solutions for practical domains. These insights support policy design, research funding strategies, and long-term planning in broader contexts of digital transformation and technological governance.

This study explores how mahallas – traditional neighborhood organizations in Uzbekistan – mediate community interventions through dense webs of social relationships shaped by cultural norms. By combining social network analysis (SNA) with ethnographic fieldwork in two urban mahallas, the research investigates how everyday practices such as hashar (collective labor) and targeted resource distribution operate as culturally grounded, naturally occurring network-based interventions.

The findings show that social cohesion, moral legitimacy, and embedded trust are central to understanding how interventions unfold and which actors become influential. While structural metrics such as degree and betweenness centrality identify key positions within the network, the ability of individuals to mobilize support or distribute aid effectively depends on culturally specific notions of authority, such as halollik (honesty), uyat (social shame), and birlik (unity). The study reveals that mahalla-based networks are not neutral delivery channels but culturally saturated systems where interventions are interpreted, adapted, and contested. Interventions that align with community values and provide opportunities for inclusive participation, such as participatory clean-up campaigns, can activate new ties, strengthen peripheral actors, and foster more equitable forms of engagement. In contrast, top-down or culturally disconnected efforts may reinforce exclusion or have limited relational impact.

This research contributes to expanding the conceptual scope of network-based interventions by illustrating how informal, community-initiated actions can produce structural and symbolic change. It also offers methodological insights into integrating network metrics with qualitative data to capture the cultural logics that animate relational life. The findings underscore the importance of designing SNA-informed interventions that are not only structurally sound but also culturally resonant, particularly in settings where governance is deeply embedded in social norms and moral expectations.

This study employs social network analysis to examine how positively energizing leadership behaviors, or behaviors that include virtuous and relational energy as perceived by followers, flow through organizational networks and impact employee well-being. While previous research has established that positively energizing leaders create vitality and enthusiasm among employees through specific behaviors like expressing gratitude and demonstrating genuine care, little is known about how these energy dynamics operate at the network level within organizations.

Using sociometric survey techniques, we will map complete organizational networks within 2-3 Irish organizations. Data collection includes perceived Positively Energizing Leadership Scale (PELS) behaviors, relational energy, vitality, and PERMA+4 well-being scores (covering positive emotions, engagement, relationships, meaning, accomplishment, plus physical health, mindset, environment, and economic security), and energy network connections.

Our analytical approach examines network centrality, density, and clustering to identify how positively energizing leadership spreads through organizational structures. We investigate whether individuals functioning as "energy hubs" (high in-degree centrality for PEL nominations) demonstrate different well-being profiles compared to energy "receivers" or "isolates." Additionally, we explore how formal leadership positions align with informal energy leadership roles within the network structure.

Using exponential random graph modeling (ERGM), we identify organizational factors that facilitate or hinder the diffusion of positively energizing leadership. This approach reveals not just individual-level leadership effects but how relational energy circulates within organizations as a collective phenomenon, potentially creating "energy hotspots" or "energy deserts" that significantly impact organizational well-being. We hypothesis that people in non-leadership roles can have influencing and positively energizing effects on networks.

Preliminary findings suggest that positively energizing leadership creates measurable network effects beyond traditional hierarchical structures, with implications for organizational design and leadership development. This research contributes to understanding how energy-based leadership constructs operate through social networks and provides practical insights for fostering positive organizational climates through strategic network interventions.

Population ageing poses significant challenges for societies, particularly in the provision of services and support for older adults. While formal services such as home care and volunteering help meet some of their needs, informal social support, provided by relatives, friends, and neighbours, are equally important. In 2020 during the COVID-19 pandemic, 7.8% of elderly individuals in Slovenia reported having no informal social support, which was a sharp increase from 2000, when less than 1% of respondents reported lacking any informal support. Research has shown that certain types of informal support networks (especially, but not limited to, those with very low number of social supporters) are less effective, especially in times of crisis, such as the COVID-19 pandemic. In 2020, overall, 42% of respondents were categorized as socially vulnerable from the perspective of social support network types.

This presentation explores changes in social support networks five years after the pandemic. Using clustering of symbolic data, we identified distinct network types based on the number and characteristics of support providers (such as relationship type, spatial proximity, and frequency of contact) across emotional, instrumental, and socializing dimensions. Findings show a partial recovery: the proportion of elderly individuals without any social support dropped to 4.4% in 2025, and the share of those with vulnerable network types decreased to 37%. These results illustrate the gradual recovery of informal social support, while also pointing to areas where older adults remain at risk of social isolation (especially men, those with lower education and income, and those from rural area).

Since the beginning of the Russo-Ukrainian War in 2022, national debates about security in general and cybersecurity in particular have become more frequent across European countries. In Italy, this debate has also been fostered by the newly elected government of Giorgia Meloni, which started operating exactly in 2022. Topics such as "national security" and "protection" towards cyber attacks - coming from outside the national borders - have begun to dominating the political arena. However, while policymakers overall agree on strengthening national cybersecurity, they are not aligned when it comes to the policies that must be implemented to achieve this objective. For instance, the recent discussion about establishing a partnership between the Italian government and Starlink has been paused because of potential security issues. In this research, we use Discourse Network Analysis to map the discourse networks among Italian policymakers discussing cybersecurity-related topics. Our objective is to identify those coalitions that may be able to support the implementation of specific policies - and infer what may be the next steps in an evolving debate that is particularly sensitive for policymakers, organizations, and individuals.

Decentralised Finance (DeFi) is a rapidly growing area of the digital economy that offers financial services without relying on traditional intermediaries. Built on blockchain technology, DeFi provides transparency and security through cryptographic protocols and distributed systems. However, it faces a core security paradox. While blockchain infrastructure is designed to provide robust technical security, users often encounter significant challenges when interacting with decentralised systems. These challenges usually stem from complex interfaces, unclear risk communication, and limited user support. As a result, users may become confused, make poor decisions, and incur financial losses.

This paper adopts Socio-Technical Systems (STS) Theory to explore how trust in DeFi platforms is shaped by both technological design and user behaviour. It focuses on the interaction between users and security interfaces such as digital wallets, transaction confirmations, and audit dashboards, critical points where users make decisions based on their understanding of risk and design clarity.

The presentation investigates how trust is constructed through these interactions and the broader ecosystem of information within DeFi communities. Adopting a Social Network Analysis (SNA) perspective, the study examines how trust-related information flows through user networks, how actors like validators, developers, and opinion leaders shape perceptions, and how network structures influence the spread of security narratives. Planned qualitative interviews with DeFi users of varying experience levels will support this analysis by offering insight into how users interpret and share trust-related information.

The presentation posits that trust in DeFi is not solely driven by technical design but also shaped by social engagement and networked interpretation. A proposed framework will guide future empirical research into how users understand and respond to DeFi security.

This study investigates the role of social networks in southern Italian students’ decision to relocate to central and northern Italy for higher education. Indeed, student mobility in Italy is predominantly a unidirectional phenomenon, taking place almost exclusively from the South to the Center and the North, further exacerbating existing regional inequalities. While the academic literature has identified key drivers of student migration, such as academic performance, social class of origin, labor market conditions, and university quality, the role of social influence has remained underexplored. To address this gap in the existing literature, a survey was administered to 209 high school students in their final year from two high schools in a southern Italian city. Social network influence was assessed using personal network analysis, and logistic regression models were computed to assess the impact of network characteristics on students' mobility decisions, while controlling for a comprehensive set of personal and contextual factors. The findings reveal a significant association between students’ mobility choices and social network measures, with the most critical factor being the proportion of their social contacts who had pursued or intend to pursue similar education and mobility paths. Noteworthy, the inclusion of social network influence measures diminished the impact of traditionally considered socio-demographic determinants, suggesting that while such factors may create the conditions for social ties to develop, it is the networks themselves that provide the mechanisms through which students adopt similar educational trajectories, thereby increasing the likelihood of migration.

The present contribution investigates how informal peer networks within high schools shape students’ soft skills, a dimension increasingly recognized as essential for educational attainment and labor market integration. Drawing on a novel whole-network survey of final-year students sample across 28 high schools in the Campania region of Southern Italy, we explore the role of multiple types of peer ties—friendship, advice, emotional, and appraisal support—in affecting core soft skills, such as autonomy, collaboration, empathy, and openness. Theoretically anchored in social network analysis and the concept of schools as relational ecosystems, the study adopts a multiplex network perspective to capture how various forms of peer interaction contribute to socio-emotional development. Using network autocorrelation models (NAMs), we analyze the strength and direction of social influence across different relational dimensions, controlling for gender, parental education, and school type.
Preliminary results reveal heterogeneous peer effects across soft skills and tie types. Collaboration is positively reinforced across all types of peer relationships, supporting social contagion mechanisms. Empathy is influenced by friendship and emotional support networks, while openness is shaped by appraisal ties, suggesting that evaluative social exchanges promote open-mindedness. On the other hand, autonomy displays a negative effect within advice and appraisal networks, indicating inverse dynamics and potential differentiation processes among peers. These first findings underscore the importance of understanding how specific social interactions within classrooms affect soft skills development. Implications for policy and practice include promoting peer-informed interventions and designing school environments that foster constructive relational ties. While the cross-sectional nature of the data limits causal inference, the research provides a strong empirical foundation for future longitudinal and comparative studies of peer effects on individual outcomes.

Adolescence is a critical period in which peer relationships strongly influence identity formation, educational aspirations, and life trajectories. While previous research has emphasized the impact of family background and individual factors on students’ academic choices, recent sociological studies highlight the central role of social networks in shaping these decisions. This study investigates how structural dynamics of peer relations and motivational homophily—the tendency to form friendships with individuals who share similar post-secondary aspirations—affect educational pathways during the transition from secondary school to university. Using Exponential Random Graph Models (ERGMs), we analyze classroom friendship networks to assess the extent to which ties are shaped by density, reciprocity, and homophily in both socio-demographic and motivational attributes. Each class network is modeled separately, and estimated parameters are subsequently aggregated through multilevel meta-analysis, enabling the examination of how contextual characteristics at the school and provincial level contribute to variations in relational patterns. By linking micro-level peer interactions with meso-level class structures, this approach allows disentangling endogenous network mechanisms from exogenous influences, thereby providing a comprehensive understanding of how peer relations contribute to educational decision-making. Findings are expected to advance sociological knowledge on the interplay between motivation, social capital, and structural opportunities, offering insights into how schools function as organizational contexts that shape students’ future trajectories.

Understanding whether and how different patterns of social media (SM) use relate to different users’ wellbeing outcomes has become an urgent academic and public-health issue. This is especially relevant now, as many countries are evaluating and implementing school smartphone bans and age-related restrictions aimed at reducing potentially harmful effects of digital technologies on youngers’ mental health.
While representing a rapid response to growing public concerns, the complexity of the phenomenon must not be overlooked. User-specific online behaviours, gender differences, developmental and contextual factors, and the role of socialization agents (e.g., parents, educators, peers) are crucial in shaping adolescents’ experience with digital technologies.
In this setting, the adoption of data-driven methods allowing to either uncover subgroups of users with similar patterns of online behaviours while simultaneously modelling data dependence structure can contribute to either advance academic research providing insights for the development of concrete intervention strategies.
To this end, we develop and administer a survey to 1155 students aged 12 to 21 years (M = 15.62) from middle and high schools in Lombardy aimed at assessing social, psychological and behavioral dimensions related to SM use.
Data were analyzed in two stages: (i) at first we applied a Graphical Mixture Model (GMM)-based clustering approach identifying latent homogeneous subgroups of users with distinct network structures; then (ii) we estimated Bayesian Networks to explore interrelationships among SM use indicators and the cluster membership derived in the first stage, performing also “what-if” analysis.
A four-cluster solution emerged as best, with clusters displaying peculiar patterns of online self-image management, psychological well-being, and perceived social support.
Going beyond overly simplistic positive/negative dichotomy, holistic statistical approaches are needed to integrate data related to different domain allowing to identify strategies that enhance adolescents’ awareness of digital tools thus promoting more adaptive online behaviors.

Research in social psychology consistently shows that men and women differ markedly in their selectivity when choosing partners for casual relationships, with women typically exhibiting more stringent criteria than men. While this robust phenomenon has traditionally been explained through evolutionary or sociocultural frameworks, we propose an alternative mechanism based on network dynamics. Our explanation derives from two fundamental properties of heterosexual matching networks: first, each match must involve one agent from each group (men and women), and second, changes in selectivity within one group directly affect matching opportunities in the other. We demonstrate how these properties create a feedback loop that amplifies any small inherent differences between the two sexes, inevitably driving one group toward high selectivity and the other toward minimal selectivity. This dynamic renders largely irrelevant within-group variation both in relationship goals and attractiveness: even an attractive man who prefers fewer, quality encounters is driven to become non-selective. This mechanism explains observed sex differences in mate selection without requiring evolutionary adaptations or sociocultural forces, though it remains compatible with their influence.

The Stochastic Block Model (SBM) is a foundational tool in network analysis, often extended to address complex problems in various domains. In this work, we develop a Bayesian network model based on an extension of the SBM where the response is categorical and denotes different type of connections between nodes. The data are represented by a large table which is similar to a contingency table but now interest lies to finding similarities in the connections between nodes. The method can be used for either sparse or dense networks without loss of generality. We use the simple multinomial-Dirichlet conjugate Bayesian model for the estimation of the model parameters and the reversible jump algorithm for the identification of blocks/clusters/communities with similar connection properties.

The proposed methodology can be used to evaluate competitive balance between teams in a sports league. We represent the outcomes of all matches in a football season as a dense network, where nodes correspond to teams and the categorical edges reflect the results of each game—win, draw, or loss.
This model is then applied to assess competitive balance, a topic of great interest in sports Economics and of general public. The primary focus of this application is on the English First Division / Premier League, covering over 40 seasons. Our analysis indicates a structural shift in competitive balance around the early 2000s, transitioning from a reasonably balanced league to a two-tier structure.

What determines a song's success at the Eurovision Song Contest? Is it driven by individual taste, geopolitical alliances, or do shared musical structures shape audience preferences? This work uses the Eurovision Song Contest as a natural laboratory to explore the complex interdependencies between music, culture, and politics through the lens of Network Analysis.
Using a purpose-built dataset, we construct two types of networks: undirected similarity networks connecting songs based on musical, lyrical, and performative features, and bipartite networks linking voting countries to the songs they voted for. We apply centrality measures and clustering algorithms to uncover structural patterns within these networks. Our findings reveal that winning entries are often not the most central in the similarity networks. Instead, they tend to occupy peripheral positions, distinguishing themselves through unconventional features. At the same time, recurring characteristics emerge among highly voted songs: the use of English lyrics, minor keys, medium-to-high tempo (BPM), and strong visual performances. Furthermore, we detect unexpected voting affinities between countries with limited historical or cultural ties.
This work illustrates how network-based approaches can uncover latent structures in cultural phenomena, offering an original perspective on collective behavior, aesthetic preferences, and strategic alliances in an international contest.

In recent years, Digital Humanities (DH) has gained significant visibility both within academia and in public discourse, yet this expansion has been accompanied by persistent debates over the field's identity and intellectual foundations (Hockey, 2004; Nyhan, 2016). This ongoing "soul searching" has produced a steady stream of publications centered on recurring questions such as "What is DH?" (Piotrowski, 2024). In this context, also due to a lack of institutional recognition, informal networks and communication channels have played a key role in shaping the community. In particular, the Humanist mailing list (founded by W. McCarty in 1987) has served as a central hub for exchange and debate.

This study applies a network-analysis approach, adapted from studies of character-word relationships in Homer's Iliad (Rydberg-Cox, in print), to the complete corpus of messages collected in the Humanist Discussion Group archive (1987–2020). The analysis requires systematic preprocessing of the textual corpus, including tokenization, vocabulary filtering, and standardization to extract meaningful linguistic patterns from unstructured email data. We construct a bipartite network where nodes represent message authors and the words they use (excluding stop words and terms occurring in fewer than a specified threshold), and edges connect authors to their vocabulary. Additionally, we examine a complementary network of direct interactions based on quotation patterns within email threads, where authors engage explicitly with each other's contributions.

We explore how this dual network approach can mine large textual corpora to identify central terms that define the DH lexicon, examine the centrality of concepts, investigate the evolution of their popularity over time, and extract latent thematic structures and semantic relationships within the DH discourse. The interaction network provides insights into actual scholarly dialogue patterns and collaborative relationships, offering a more complete picture of community formation and intellectual exchange.

This contribution focuses on the study of scientific literature in the field of Bioregenerative Life Support Systems (BLSSs), which is at the forefront of current space science. Since the 1960s – and more intensely since the 1980s – this multidisciplinary field has brought together scholars from various disciplines and different countries to investigate how to sustain human life over long periods in extreme environments, such as those of outer space. BLSSs can provide space crews travelling to or settling on the Moon or Mars with water, oxygen and food, and are also designed to recycle waste and regenerate air. The complexity of these tasks implies that different sciences must cooperate to develop BLSSs as hybrid systems apt to fulfil vital physicochemical processes. In fact, BLSSs open up new ecological perspectives thanks to the encounter between engineering, biochemistry, biology, chemical engineering, ecology, cybernetics, physiology, medicine, and agricultural science. As such, multidisciplinary efforts in BLSS research deserve bibliometric network studies that help to unveil the patterns of collaboration among the relevant scholars and institutions. It is also important to understand how innovations in BLSS research can foster future expectations towards achievements in self-sustaining closed ecological systems and the related advancements in the production of scientific knowledge. Furthermore, BLSS research is characterized by a dual perspective, seeking to understand how to live in both outer space and extreme conditions on Earth and thus aiming to improve our knowledge of earthly ecosystems. The current proposal is based on co-authorship and co-occurrence analysis of the relevant literature over a temporal dimension spanning more than forty years, showing how the field has grown and the extent to which differentiation and association among diverse topics, disciplines and scholars characterise this multifaceted scientific domain.

There is a strong consensus in the social and health sciences that the quality, quantity, and diversity of social connections are vital for human well-being. Social relationships have intrinsic value and play an instrumental role in contributing significantly to physical health. As a result, disparities in social relationships not only reflect social inequality but also contribute to unequal health outcomes across different populations. This study investigates the structural and qualitative disparities in social networks among individuals, with a focus on socio-economic status (SES), urbanization levels, and migration background. We hypothesize that social disadvantages are reflected in the composition of social networks and the dynamics of support within them. Using data from the Multipurpose Survey on households of the Italian Statistical Society, we examine whether individuals with low SES, those living in less urbanized areas, and individuals with a migration background have social networks characterized by smaller sizes, reduced support provision and limited resources. To enrich our understanding of migrant populations and social integration dynamics, we complement official national statistics with insights from international and independent sources, such as ISMU, which provide valuable contextual information beyond institutional datasets. Additionally, we explore how network structure impacts health behaviors and perceptions.

One of the most significant challenges facing society is the development of effective mechanisms that can ensure the security of citizens while facilitating comprehensive economic and social advancement. This study identifies the commercial areas of Bogotá, Colombia, which are critical points in terms of security, and proposes a strategic allocation of police resources in order to optimise their coverage and response times. The methodological process was executed in accordance with the guidelines set forth in the Cross-Industry Standard Process for Data Mining. Geospatial analysis techniques were employed, including the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to identify commercial zones and the graph theory to model the security network. Additionally, data from the Special Administrative Unit of Land Registry and the locations of National Police stations and precincts were utilised. The results demonstrate the identification of 241 commercial zones in Bogotá and propose a strategic allocation of police resources based on degree centrality and a stable matching algorithm. The integration of related techniques allowed for the generation of an innovative approach that combines the accurate identification of critical commercial zones with a data-driven optimisation of police resources. This not only increases the understanding of security dynamics in urban commercial areas, but also provides a robust public safety decision-making tool for the city.

This paper examines the characteristics of inter-organisational networks in post-disaster recovery, focusing on their ability to adapt and remain effective over time. While extensive research has explored such networks during the preparedness and response phases, their role in long-term recovery has received less attention. Unlike the response phase, recovery is a multi-year process involving the sustainable restoration of the physical, economic, social and psychological well-being of affected communities.
To provide a more systematic basis for studying the sustainability of inter-organisational networks, this study conducts an integrative review, synthesising knowledge across research traditions to identify common themes.
In line with the PRISMA framework, a keyword search was conducted in major academic databases, including Scopus, EBSCOhost and Web of Science. The selection process involved screening of titles and abstracts, followed by a full-text review based on pre-defined inclusion criteria, resulting in 17 papers using a network approach. Inductive coding was combined with deductive coding based on the collaborative resilience framework, covering diversity of actors and resources, trust and reciprocity, leadership, community buy-in and structure.
The findings show that recovery networks often start out with centralised coordination models, but tend to evolve into more decentralised forms over time. However, even if decentralisation is desirable, this does not always result in a more effective recovery. In fact, it can lead to fragmentation, unclear roles and asymmetries between national and local actors.
Recovery networks evolve within hybrid governance systems, where effectiveness depends on balancing authority, resources, and local agency over time.

Girls, historically disadvantaged, outperform boys in education across most OECD countries, despite persistent gender gaps in STEM and labour markets. While boys outperform girls in standardised math assessments, girls consistently receive higher grades across subjects. This paradox has sparked growing interest in the Gender Grading Gap (GGG), defined as the advantage girls receive in teacher-assigned grades compared to boys with similar performance on external assessments. The GGG has been mainly attributed to gender differences in classroom behaviour and teachers' implicit stereotypes. This study investigates four dimensions of the GGG not fully addressed in the literature: (1) variation across school subjects, (2) the role of school/classroom context and peer composition, (3) temporal dynamics over the school career, and (4) consequences for student performance. First, we provide a systematic, theory-driven analysis of the GGG in literacy (Italian) and numeracy (Math). Second, we examine contextual moderators—such as classroom gender balance, socioeconomic composition, and migrant background—highlighting how peer effects may influence behaviour, performance norms, and grading criteria in gendered ways. We also test whether grading practices reflect compensatory/reinforcing preferences by classroom-level gender gaps in standardised tests. Third, we assess trends in the GGG across grades and cohorts. Fourth, we explore whether early exposure to gendered grading patterns predicts later performance. This study utilises INVALSI population data, encompassing over 7 million students (grades 2, 5, and 8) across 371,423 classrooms from 2012 to 2019. Longitudinal analysis follows the 2012 grade 2 cohort through grades 5 (2015) and 8 (2018). Estimation is performed separately by subject and grade at the classroom level, allowing analysis of grading heterogeneity and moderation by class gender, ESCS, and migration composition. The study finds that the GGG is larger in mathematics, shaped by classroom context, and predicts better outcomes for girls, highlighting the long-term impact of subjective grading on gender inequality.

In Italy, the establishment of online universities was officially sanctioned by the 2003 Finance Act. Despite this, academic research on the topic remains limited. Online universities are often perceived as having lower academic prestige and are commonly associated with students seeking degrees primarily for public sector employment. However, these institutions also provide significant advantages in terms of flexible scheduling and the elimination of relocation and housing costs. To contribute to this ongoing debate, we apply a propensity score matching approach to compare dropout rates between cohorts of students enrolled in traditional and online Italian universities between 2010 and 2019. Preliminary findings suggest that attending an online university has a positive effect, particularly for students who are typically at higher risk of dropping out from traditional institutions.

This contribution presents the research conducted as part of the PRIN 2022 project From High School to University: Assessing Peers’ Influence in Educational Inequalities and Performances, carried out by the research units of Cagliari, Florence, and Salerno and has the main aim to provide an insight on its two main objectives and research outcomes.
The research provides evidence for analysing the mechanisms behind the reproduction of educational inequalities from upper secondary school to university, with a focus on socioeconomic conditions, gender gaps, and peer effects. Particular attention is devoted to the role played by school and peer context and to the extent to which school contexts reduce or amplify initial disadvantages and gender disparities, acting as drivers of egalitarian processes.
As for the first objective, national-level data from the Italian INVALSI surveys in upper secondary schools are used to
assess disparities in learning outcomes at grade 13 and
identify patterns of inequality at school level. In addition, the MOBYSU.IT dataset, which contains information on students’ careers at university, linked with INVALSI data at grade 13, is employed to analyse the transition from high school to university and to monitor high schools effectiveness.
As a result, the first research outcome is a system of indicators for monitoring and mapping educational inequalities in high schools, with reference to educational choices, skill levels, and university trajectories.
For the second objective, three surveys have been carried out by the three research units in their respective territories (Campania, Tuscany & Sardinia) through the administration of questionnaires to students in the fourth and fifth years of high school. They provide evidence to shed light on factors influencing educational choices and the role of social influence in the decision to continue studying, the choice of disciplinary field, and the preference for attending a local or non-local university.

This paper explores the evolution of Italian hip-hop from the 1990s to the early 2000s by combining social network analysis, computational linguistics, and qualitative methods to investigate how collaboration structures shape cultural expression. Drawing on a dataset of tracks from several influential albums, a collaboration network of artists has been built though Gephi, revealing distinct communities that reflect both stable partnerships and dynamic cross-regional exchanges. Central artists, such as Guè Pequeno and Marracash, tend to occupy mainstream positions within the network and thematically engage with individualism, materialism, and status. In contrast, peripheral artists are more likely to foreground themes of resistance, urban marginality, and political critique. By connecting artists’ positions within the collaboration network to the thematic content of their lyrics, using TF-IDF metrics complemented by qualitative analysis, the study shows how collaboration patterns influence the type of narratives and values promoted within the genre. This approach sheds light on the broader cultural dynamics at play, revealing how Italian hip-hop navigates between local roots and global influences, and how the tension between authenticity and commercialization shapes its thematic development.

A fundamental challenge in the sociology of taste is reconciling our intuitive yet often “fuzzy” understanding of musical genres and social categories with traditional methods that depend on clear categories and central tendencies. This research addresses this issue by applying relational (dual projection methods for two modern networks) and geometric (stacked Correspondence Analysis) techniques to explore the perceived connections between musical genres and social characteristics, specifically focusing on subjective interpretations of these linkages rather than strictly objective connections. The study utilizes data from a representative sample of Americans (N = 2250), each of whom reported their perceived associations between 20 musical genres and 15 social characteristics, treating this information as cognitive two-mode data. The methodological approach involves several key steps: First, Dual Projection (Breiger, 1974; Everett & Borgatti, 2013) and Backbone Extraction (Neal, 2014) are used to identify and isolate the most significant perceived associations in the data. This "Mondo Breiger" approach generates a personal two-mode network for each individual, from which personal genre projections (the perceived similarity between genres based on shared labels) and personal label projections (the perceived similarity between labels based on shared genres) are derived. Next, these individual projections are transformed by extracting binarized backbones. The backbone of these person-specific projected networks is modeled jointly using Stacked Correspondence Analysis (CA). This yields three sets of scores: person-specific judgments of relative similarity, aggregate judgments, and supplementary scores that represent the centroid of personal judgments for genres and labels. Geometric Data Analysis is then used to examine the distribution of various groups of individuals, analyzing how they are arranged along principal axes in genre and social label spaces. This comprehensive approach shifts the focus from rigid definitions to more flexible categories, allowing for a direct examination of diversity and the distribution of opinions across the social spectrum.

Long established lines of research have discussed the relationship between social class and cultural taste. In this paper, we build on Peterson’s distinction between omnivore and univore tastes to explore the link between cultural consumption and participation in associations and cultural life in the community; in particular, focusing on members of cultural associations, we ask whether cultural capital (formal education) still influences their adoption of omnivore styles of cultural consumption, after various forms of social participation and the resulting networks are controlled for. We do so by drawing upon data collected between 2021 and 2023 among members of brass bands (N=810) and choirs (N=1876) in the province of Trento, Italy. Treating the two groups separately, we operationalize omnivorousness in two different ways: (a) as an additive measure of interest in various genres; (b) as an additive measure of “contamination”, i.e., interest in genres that a duality analysis shows to belong usually in different musical families (e.g. rap and classic music), while discounting interest in genres usually regarded as close (e.g. jazz and blues). Community life is operationalized in both categorical and relational terms. In categorical terms, we look at overall levels of participation in associations and local cultural activities. In relational terms, we look at networks originated from the duality of individuals and associations/cultural practices. We plan to run regression models which include centrality measures (betweenness and closeness) and estimates of the heterogeneity of respondents’ networks. They will enable us to expand on earlier findings, suggesting that participation in associations and involvement in networks may substantially reduce the impact of education and status over omnivorousness.

The latent position model is a popular approach for the statistical analysis of network data. A key aspect of this model is that it assigns nodes to random positions in a latent space, and the probability of an interaction between each pair of nodes is determined by their distance, allowing researchers to visualize nuanced structures via a latent embedding of the graph. Missing data is a common issue in statistical network analysis, often leading to an excess of observed zeros in interaction data. In this talk, I will focus on non-negatively weighted social networks. By treating missing data as "unusual" zero interactions, we propose a combination of the zero-inflated Poisson distribution with the latent position cluster model. Our framework extends the latent position model to accommodate the clustering of individuals and to simultaneously model weighted interactions, handle missing data, perform clustering, and produce three dimensional visualizations of real networks. Statistical inference is based on a partially collapsed Markov chain Monte Carlo algorithm, which involves a new truncated absorb-eject move, and selects the number of groups automatically by leveraging a mixture of finite mixtures framework.

Longitudinal social network data, observed as discrete-time snapshots, can be modeled using the dynamic stochastic block model, which assumes that nodes are clustered into unobservable groups and that block memberships evolve according to a hidden Markov chain. This model can be used as an early warning system by predicting the occurrence of future links through the evolution of latent block memberships. Considering the sequence of binary adjacency matrices $\mathbf{Y}^{(t)}$, $t=1,\ldots,T$, with entries $Y_{ij}^{(t)}=1$ if nodes $i$ and $j$ are connected by an edge at time occasion $t$ ($i,j=1,\ldots,n$), the model relies on node-specific discrete latent processes $U_i^{(1)},\ldots,U_i^{(T)}$, whose parameters are the initial probabilities $\alpha_u=p(U_i^{(1)}=u)$, the transition probabilities $\pi_{u|v}=p(U_i^{(t)}=u|U_i^{(t-1)}=v)$, and the connection probabilities $\beta_{uv}=p(Y_{ij}^{(t)}=1|U_i^{(t)}=u,U_j^{(t)}=v)$. Estimation is carried out using an approximated maximum likelihood approach based on a variational approximation of the log-likelihood function.

We propose a novel link prediction method to forecast possible connections between nodes $i$ and $j$ at time $T+1$ by estimating the following edge probability
\begin{equation}
p_{ij}^{edge}=\beta_{\hat{u}_i^{(T+1)},\hat{u}_j^{(T+1)}},
\end{equation}
where $\hat{u}_i^{(T+1)}=\arg\max_{u=1,\ldots,k}\hat{q}_i^{(T+1)}(u)$ is the predicted block for node $i$ at time occasion $T+1$ computed on the basis of the maximum a-posteriori rule. Here $\hat{q}_i^{(T+1)}(u)$ is the variational approximation of the conditional posterior probability that node $i$ belongs to latent block $u$, predicted for time point $T+1$, given the observed network.
A connection is then predicted when $p_{ij}^{edge}$ exceeds a certain threshold, which can be determined based on F1 score.

The approach is validated through simulations across a wide range of scenarios and it is applied to public data about interactions among 164 ants observed over eleven days. Accurate predictions are obtained especially when the network is dense and relatively persistent.

We propose a dynamic extension of the Mixture of Latent Trait Analyzers (MLTA) for a bipartite network. Specifically, we move along a Hidden Markov Model framework to account for the dynamic nature of the data and enable a dynamic clustering of sending nodes over time. A multidimensional continuous latent variable (trait) is assumed to account for residual, unobserved, time-constant, latent factors that may affect the way sending nodes relate to the receiving ones. Estimation of model parameters is conducted within a maximum likelihood framework by extending the Baum-Welch algorithm to account for the presence of the continuous latent trait in the model. However, this algorithm requires the solution of multidimensional integrals over the latent trait domain that are not available in closed form. To overcome the issue, a variational approach is employed. This consists of approximating the intractable integrals with lower bounds that are easy to manage and solve. The effectiveness of the proposal is demonstrated via an extensive simulation study, based on a varying number of sending and receiving nodes, as well as a varying number of time occasions. The proposal is also employed for the analysis of data from the Survey of Health, Ageing and Retirement in Europe (SHARE) with a focus on Italian residents. The aim is that of dynamically cluster residents according to their mental health status, while also accounting for possible unobserved latent factors related to their psychological well-being.

This paper offers methodological guidelines for the application of blockmodelling (BM), a clustering technique that historically informed heterodox analyses of trade but has since fallen out of favour, to the internataional trade network. It also puts these recommendations at work in a two-snapshot longitudinal case study into the the transformation of international trade under the first Trump administration (2017-2021). Namely, the paper examines the extent to which shifts in US trade policy (trade wars, renegotiated agreements, emphasis on economic nationalism) altered the structure of the ITN in those years.
The analysis compares existing BM approaches and assesses their suitability for capturing heterogenoeus trade patterns amidst exogenous shocks such as tariffs and sanctions. Furthermore, this study addresses critical methodological challenges relating to the ITN's scale-free properties and heterogeneities in relational capacities such as data normalisation. It also discusses the potential to integrate insights from the gravity model of trade, a standard econometric model, into BM. By bridging social-network analysis and international trade, this study both revitalises BM as a valuable tool in international economics and provides an empirical reassessment of the earlier 'Trump effect'.

Kin is the most important source of social support and wellbeing across societies. Human kin networks vary in the presence and availability of living kin and are influenced by changes such as the ongoing demographic transition. This refers to long-term changes in fertility and survival rates in developed and high-income societies. One result of this demographic change is a narrowing of the available kin network over an individual's life course. Kin networks become more vertical, meaning they primarily consist of relatives from direct lines of descent. However, significant differences in kinship composition by kin type between individuals are likely to continue in such populations.

The Social Networks, Fertility and Wellbeing in Ageing Populations (NetResilience) consortium investigates demographic change from the perspective of social networks. NetResilience was launched in 2021 and will continue until 2027. So far, we have studied kin effects and kin networks in historical Finland, kin effects on fertility and wellbeing in contemporary societies, and many other topics related to social networks and wellbeing (see consortium publications: https://www.netresilience.fi/en/publications/).

The current presentation will briefly introduce our ongoing work with kinship data compiled from Finnish register data, which covers the entire Finnish population. We aim to examine the influence of the existence (that is, never born, alive, or deceased) of both biological and affinal kin up to first cousins on the survival of an index individual (referred to as "ego"), using complete population register data and kin networks from Finland. The idea is to apply machine learning techniques in a data driven way to identify which relatives, and at which points in life, are most beneficial for longevity.

The rise of remote work has introduced a new axis of inequalities in the labor market, reshaping both where and how people work. While research on social networks has long emphasized the importance of personal ties in job searches and status attainment, the rapid diffusion of remote work—now nearly four times pre-pandemic lev-els—opens new and largely unexplored avenues of inquiry. Drawing on egocentric network data from 4,097 workers and 22,402 professional and household ties, we examine how three distinct work arrangements—fully remote, hybrid, and on-site—shape the structure and quality of personal networks. Using multivariate and multi-level techniques, we test hypotheses focused on access to and mobilization of social support, the emergence of workplace and household conflict, and structural inequalities. Our findings show that fully remote workers have more geographically dispersed networks and, overall, less conflictive relationships with colleagues in lower or equal hierarchical positions. In contrast, managers working remotely report significantly higher proportion of difficult ties. Moreover, on-site work appears to shield against the emergence of tensions with children. This article provides novel empirical evidence on a growing and powerful driver of personal network dynamics in today’s labor market.

As teleworking persists beyond the pandemic, it continues to reshape individuals’ daily routines and social interactions. While teleworking reduces face-to-face contact with colleagues, it eliminates commuting time and offers greater flexibility, potentially increasing engagement with family and friends. However, few studies have applied a personal network approach to understand how teleworking restructures everyday social ties.
In this study, we use personal network data to examine how teleworking shapes the frequency of interactions across different types of close ties, as well as the compositional and relational characteristics of individuals’ ego networks. We also assess whether these patterns are moderated by personality traits, drawing on the social dimensions of the Big Five model.
Utilising data from the Dutch Longitudinal Internet Studies for the Social Sciences (LISS) panel, we compare teleworkers and commuters in terms of contact frequency and core personal network composition. We find that, contrary to expectations, teleworkers report significantly smaller personal networks, though they maintain higher contact frequencies, especially with family ties. Personality traits are directly associated with ego network characteristics in largely expected ways, and we find some evidence of interaction effects between personality and teleworking status in shaping social outcomes.
These findings highlight how teleworking can alter private-sphere social networks and demonstrate the relevance of individual dispositions in understanding the broader social consequences of remote work.

Previous research across multiple disciplines has documented the positive nexus between good peer relationships in the classroom and learning. Robust social networks create inclusive learning environments, enhance academic performance of more fragile students, build children’s confidence, and foster resilience in young learners. In this study, we examine how structural position in classroom’ networks influence mathematics performance among 3rd grade primary students. We hyphotize that “peer” interactions facilitate mutual support, thereby affecting academic outcomes. Moreover, we consider the relevance of student “pairing” based on different aspects of homophily/heterophily patterns.

Data were collected as part of the MATES project (https://www.progetto-mates.it/), funded by the Italian Ministry of University in 2022. A total of 3,316 students from 180 classes across 78 schools in Milano and Napoli participated in our study. At the beginning of the school year 2024/2025, students completed a questionnaire that gathered personal background information (e.g., gender, preschool attendance, family background), attitudes toward mathematics (using a psychological scale), cultural habits (e.g., sports, playing musical instruments, television viewing), and responses to ego-alter questions used to define the classroom’s overall social network.

Using standardized mathematics test scores as our outcome measure, and collected from national INVALSI procedure, we assess the influence of students’ positions within the classroom network while controlling for several individual factors. Our findings show both expected and unexpected correlations, offering insights about the not linear influence of pupils' networks on their educational outcomes.

This work presents a systematic and in-depth review of the literature on peer effects, with the aim of providing a clear and up-to-date overview of the theoretical and methodological developments that characterize this field of research.
Firstly, the analysis focuses on the main theoretical frameworks, with particular attention to reference theories and studies on social interactions in educational contexts. These perspectives allow us to understand how peer relationships influence the formation of aspirations, motivation, and, more generally, educational choices. Secondly, the review examines the different methodological approaches used to operationalize these concepts, highlighting the distinction between individual-level indicators—such as average peer performance—and measures based on the structure of social networks, derived from network analysis. This dual perspective has enabled the literature to develop increasingly advanced statistical methods for studying peer effects.
Finally, the review identifies several significant gaps, particularly regarding peer effects on educational choices, a research area that remains underexplored.

This contribution examines inequalities in student mobility trajectories in Italian higher education before and during the COVID-19 pandemic, adopting a regional perspective. Italy represents a significant case, marked by persistent territorial disparities and long-standing oneway flows from Southern regions towards Northern and Central areas. Using cohort data from the Italian Student National Archive, we construct complex network data structures and apply community detection algorithms and regression models to explore how the pandemic reshaped internal mobility flows, altering established patterns. Findings show that internal mobility was affected: Northern universities remained nationally attractive, though slightly less than in the pre-pandemic period, while new routes towards Central regions emerged, signalling alternative educational opportunities. Moreover, the enduring appeal of STEM programmes is confirmed, underscoring persistent field-specific inequalities.

Analyzing co-authorship relationships provides valuable insights into collaboration patterns and their impact on research productivity and performance. This study focuses on the co-authorship relationships among academics affiliated with Italian universities in the fields of statistics, management, and sociology.
We compiled a comprehensive list of Italian scholars and their attributes, including gender, role, academic sector, university, and department, using data from the Ministry of University and Research. After linking each scholar to their Scopus ID, we collected co-authorship data by downloading all their publications from the Scopus online bibliographic archive, covering the period from 2012 to 2022. We use an ego-network approach to analyze the scientific collaboration networks in which each Italian scholar is embedded. To characterize personal networks, we employ structural and compositional indices, such as component ratios, density, and diversity, and apply network clustering to identify different types of networks. Additionally, we investigate interdisciplinary collaboration and its effects on knowledge production and dissemination.

In recent years, Research Infrastructures (RIs) have assumed a strategic role in the development of the social sciences, functioning not only as technical platforms for data collection and management but also as epistemic actors that shape scientific practices, methodological standards, and forms of collaboration. Aligned with the FAIR principles and the Open Science framework promoted by the European Union, and in line with the priorities identified by the European Strategy Forum on Research Infrastructures (ESFRI), RIs facilitate the production of interoperable, sustainable, and accessible data, thereby reducing fragmentation and enhancing the impact of research on society.
Within this context, the FOSSR project (Fostering Open Science in Social Science Research) aims to build an infrastructure for the social sciences, including the Italian Online Probability Panel (IOPP) – the first probabilistic online panel representative of the Italian population; alongside the implementation of established international surveys such as the Survey of Health, Ageing and Retirement in Europe (SHARE), the Generations and Gender Programme (GGP), and Growing Up In Digital Europe (GUIDE).
This study investigates the role and evolution of RIs through a bibliometric approach. The analysis reconstructs the conceptual, intellectual, and social structures characterizing scientific production linked to RIs, highlighting both the dynamics of quantitative growth and the trajectories of thematic development. The results indicate a shift from the predominantly technical and engineering origins of RIs, associated with topics such as Spatial Data Infrastructures and GIS systems, toward a field characterized by work carried out in socio-humanistic areas. RIs thus emerge as central nodes in the transition toward a more open and collaborative social science.

Much effort has been made to analyze industrial districts (IDs) because of their local influence on both employment rates and economic growth. IDs are rooted in locally defined environments, but they are also embedded in global value chains that influence their strategic approaches and their capacity to generate effects on local communities. Companies operating in IDs, which are mainly family firms, are particularly sensitive to significant and breakthrough events (e.g., organizational succession, financial operations, technological transformations). Among these, organizational succession significantly impacts the way they interact: as members of districts, companies rely on interrelated activities to remain innovative and generate value. Failing in organizational succession may have negative effects both on the ID itself, with the risk of its decay (e.g., skill and knowledge dispersion), and on the company’s internal structure, hindering its ability to preserve its innovation capabilities. Studying the inter- and intra-organizational relational aspects characterizing IDs and companies can shed light on their ability to address external shocks. In this study, I will adopt a combination of systematic literature review and bibliometric analysis to explore recent literature developments and provide an overview of their evolution. In particular, time-defined data will be collected from the Scopus database and then analyzed through two specific bibliometric tools. First, a co-word analysis will be conducted on specific keywords to determine conceptual and semantic maps. Second, a bibliographic coupling on shared references will be used to identify actual and emerging streams of research. The study will focus on the unexplored streams of social impacts of organizational successions, both on the social relations of IDs and inside companies, highlighting how the organizations at both levels react and adapt.

This study traces a decade of collaboration among Italian scholars across four macro sectors-two in economics and statistics and two in political and social sciences. It examines how formal collaborations, retrieved from Scopus, and informal collaborations, gathered through an ego-network questionnaire, influence academic performance. Drawing on Scopus records of all co-authored publications within these four macro sectors, published between 2012 and 2022, we can accurately map established networks. The questionnaire, distributed to scholars identified through these records, captures the informal side of collaboration.
The questionnaire is administered using a computer-assisted web interviewing (CAWI) approach via M.a.Sco.Net, a web application designed specifically to collect ego-centered network data. The web-based questionnaire comprises two sections:
1. The first evaluates both the formal and informal collaborations that have led to publications (one page per co-author);
2. The second measures informal ties with colleagues who have never co-published, mapping broader professional networks.
Within the web app, the researcher can choose between two options: either show respondents the full list of their co-authors or a subset selected by the researcher. If only a subset is shown, once the questions for those names have been completed, the app allows the respondent to add one co-author they consider most representative and to answer the same questions for that person.
Egos were sampled using a stratified design-based on scientific-disciplinary sector, academic role, gender, and co-authorship activity ( 5 vs. > 5 co-authors), drawn from MUR Population Data 2022. For each ego, up to four alters were randomly selected, with probabilities proportional to prior co-authorship frequency. Preliminary results from the ongoing investigation into how formal and informal collaborations influence academic performance in Italy will be presented and discussed.

This study explores the determinants of loneliness and perceived social isolation, focusing on the roles of personal networks and social context characteristics. Perceived social isolation is an individual’s subjective feeling of disconnection, regardless of actual social interactions. Loneliness arises from a perceived gap between desired and actual social relationships, encompassing both the quantity and quality of connections. We examine the associations between these feelings and specific characteristics of personal networks, including their structure and composition of social roles, support exchange, temporal dynamics, and modes of communication.

First, we focus on personal network characteristics as potential determinants of perceived isolation and loneliness: network size;emotional closeness;geographic distance;network cohesion;diversity of roles and interaction foci;sociodemographic diversity;and network turnout. Second, we investigate the role of online interactions, exploring whether virtual communication mitigates or exacerbates these feelings. While online platforms expand the reach of social networks, their effects on perceived connection and satisfaction with relationships remain mixed. We hypothesize that reliance on online communication weakens social bonds typically fostered through face-to-face interactions, potentially leading to a greater sense of isolation and loneliness. Finally, we consider whether the effects of personal network characteristics on perceived isolation and loneliness vary by gender, race/ethnicity, and migration status.

Using three waves of egocentric network data from the UC Berkeley Social Network Study (UCNets), which provides detailed longitudinal measurements of personal networks and individual-level observations, this study makes both theoretical and empirical contributions. First, it refines and advances theories about social and relational determinants of loneliness. Second, it provides new insights into how modern modes of social interaction, particularly online communication, influence negative feelings associated with social interactions. Ultimately, this study enhances understanding of the complex interplay between social context, individual characteristics, and modes of interaction in shaping experiences of loneliness and social isolation.

We aimed to depict, using data from the SHARE COVID-19 project carried out in 2020 and 2021, the personal interaction of elderly people in Italy, with a particular focus on tracing the changes related to the outbreak of the SARS-CoV-2 virus. To do that, we first provided face-to-face and virtual network types observed by living arrangements of elderly people and waves of the SHARE Corona Survey (SCS). Exploiting the longitudinal nature of SCS, we looked at the transition or consistency of older people in their network types during the pandemic and post-pandemic period. Two different models for the probability of observing a network-type transition have been proposed to evaluate if elderly individual characteristics (sociodemographic) or pandemic-related behaviors and conditions are associated with the changing of the network type. The percentage of older adults without face-to-face contact with others (the No Alters network type) was very high during the most challenging pandemic, then strongly decreased in 2021. Virtual contacts played a role in partially preserving some relationships, especially familiar ties. Overall, individual conditions related to sociodemographic and economic vulnerability were clearly recognized among people who remained in the network type they were embedded during the first outbreak. Living in rural areas emerged as a favorable context for resuming social contacts.

Even though HIV incidence is falling globally, it remains relatively high amongst South African youth. Despite the availability of free and effective prevention and treatment methods, uptake remains low. Evidence suggests that sexual behaviour is strongly affected by social connections and interventions uptake can be improved by leveraging social networks. We therefore use social network data from three rural communities in KwaZulu-Natal, South Africa, to evaluate how social processes affect youth’s sexual behaviours and attitudes. To do so, we interviewed cca 1000 16-30 years olds about their sexual health, behaviours and attitudes, as well as support networks, which we then connected into three community based sociocentric networks. Using auto logistic actor attribute models (ALAAMs), we investigate how HIV and Herpes Simplex (HSV-2) statuses, misbeliefs and knowledge about sexual health, capability to follow safe sexual practices, and risky sexual behaviours are patterned across these networks. Even though our results show evidence of clustering and simple contagion across all variables, they appear to follow different patterns and strength, suggesting the presence of different underlying network mechanisms behind their diffusion. These findings can be used to guide the design of more targeted and effective interventions to improve the sexual health of young South Africans.

In recent years, the literature has proposed different studies that combine the analysis of node-level attributes alongside topological information of the network. These proposals range from hierarchical clustering algorithms including relational constraints to communities in the context of Subgroup Discovery, and data-driven probabilistic methods on multilayer networks. While these methods have been proposed to integrate structural and attribute-based information, achieving a balanced and coherent representation remains challenging. In this work, we apply DISTATIS, a three-way multidimensional scaling technique, to jointly analyze network topology and node attributes. Through simulations on networks with different attribute types, our results demonstrate that DISTATIS effectively captures the coherence between the attributes (qualitative and quantitative) and network structure. This approach offers a valuable tool for extracting complex relationships in real-world networks where both structural and attribute-driven factors are crucial.

This study investigates the role of structural holes and islands in the dynamics of scientific knowledge production within social science methodology. Structural holes are conceptualized as gaps in knowledge due to fragmented connections between thematic groups, while islands represent cohesive clusters of specialized knowledge. Employing a bibliometric and scientometric approach, we analyze co-occurrence networks derived from publications to map dominant themes and identify broker-nodes that bridge isolated knowledge clusters. These brokerage nodes are hypothesized to facilitate interdisciplinary connections and foster the emergence of innovative scientific themes. Using a modified version of Cobo’s Strategic Diagram, informed by network topology metrics such as clustering and closeness centrality, we demonstrate that bridging structural holes enables the discovery of latent concepts and the evolution of scientific fields. Our method highlights traditional, innovative, and potentially innovative themes, offering a complementary tool to existing bibliometric analyses. While applied here to social science methodology, this approach holds promise for broader disciplinary applications in understanding and fostering knowledge development.

The study explores the evolution of a dynamic co-authorship network among European countries between 1990 and 2023 arising from the interplay of geographical, linguistic, and historical factors to shape scientific collaboration. Using bibliometric data from OpenAlex, Indirect Blockmodeling and Dynamic Stochastic Blockmodeling (DSBM) are employed in the analysis to examine patterns in three critical periods: rise of the Internet (1994–2003), enlargement of the EU (2004–2013), and the European Research Area (ERA) initiatives (2014–2023). The findings reveal exponential growth in the number of co-authored publications, an overall increase in intra-cluster collaboration, notably in the Balkan, Scandinavian, and Western clusters, coupled with persistent regional disparities. Despite EU policy interventions, collaborations between Western and non-Western regions remain limited. The study underscores the need for targeted measures to foster more inclusive and balanced scientific networks across Europe.

Addressing the boundary-specification problem in social network research, this paper reconstructs the collaborative network catalysed by Italy’s Impresa 4.0 policy framework, focusing on enterprise digitalisation initiatives.
The protocol was first tested in three Italian regions—Emilia-Romagna, Veneto and Piemonte—and has since been applied in five others, providing further evidence of its portability. The observation window was delimited territorially (initiatives within each region), temporally (January 2020 to March 2023), and relationally (publicly documented partnerships). Data were collected following a theory-driven, two-step procedure. Firstly, we identified the policy-mandated knowledge intermediaries—Competence Centres, Digital Innovation Hubs, and Chamber of Commerce Digital Enterprise Points—from the national Atlante I4.0 database. We then recorded all digitalisation initiatives they promoted and the partner organisations involved in each. Secondly, we identified the employer associations that appeared in at least one of these initiatives, mapped the additional projects they promoted, and recorded their associated partners. Employer associations were selected for their recognised intermediary role in the governance of the Impresa 4.0 policy; restricting the network expansion to this specific actor category preserved conceptual coherence and prevented the uncontrolled growth typical of full snowball sampling.
The resulting actor–event matrix interweaves initiatives led by policy intermediaries with those led by connected employer associations. From this two-mode network, a weighted one-mode projection yielded two distinct boundary scenarios—a 'policy-only' view and an 'expanded' view—whose topologies can be systematically compared both within and across regions. Application to the additional regions confirmed that the protocol maintains analytical clarity whilst remaining sensitive to region-specific governance arrangements.
By combining explicit boundary rules with actor–event network reconstruction and one-mode projection, this study provides a replicable template for evaluating regional policy networks with fragmented data. This approach highlights that the definition of network boundaries is not merely a preliminary step, but a substantive analytical choice that conditions all subsequent inferences.

In recent decades, the container terminal industry has experienced a significant increase in equity cooperative agreements among Terminal Operating Companies (TOCs), liner shipping operators, and transport chain operators. This development has led to a complex network of relationships that impacts national and international port-level competition. This study focuses on Italian container ports and analyzes the evolution of TOC networks at three time points (2011, 2015, and 2021) to investigate their effects on intra-port competition.
The analysis covers 25 TOCs and 191 actors, employing network statistics, visualizations, and a Stochastic Actor-oriented Model to assess the dynamics of equity partnerships—creation, maintenance, and dissolution—considering endogenous and exogenous factors such as vertical and horizontal integration, similarity in activities, and geographic distance.
The findings indicate that a few leading TOCs have increased equity ties pursuing vertical integration strategies, significantly affecting competition within Italian ports. Network visualizations highlight emerging patterns and hidden relationships, offering an innovative tool for Port Authorities and European and national policymakers to monitor potential anti-competitive behavior and monopolistic positions.
This study advances the understanding of TOC strategic behavior in the domestic market and presents a novel methodological approach to analyze competitive dynamics in container terminal management. The research aligns with European Commission policies promoting fair competition and sustainable port development.

In Network Analysis, the topological properties of networks can be investigated through subgraph analysis. In particular, interactions between three vertices can be analyzed via the so-called triad census. However, the conventional procedure is suitable only for binary (unweighted) networks, neglecting the level of heterogeneity that can be observed, particularly in small and dense networks. This paper introduces and explores the usefulness of a new algorithm, named “network peeling”, proposed to extend triad census in the case of weighted and directed networks. The proposed algorithm operates on a nested sequence of binary sub-networks in which arcs are “peeled out” each time by a unit value. Through a simulation study, we investigate whether the conventional and new weighted triad census exhibit non-negligible differences, considering three data-generating processes for the network formation, with varying density and variability of arc weights. As a well-studied case of small and dense weighted and directed networks, we examine a real-world application concerning the passing distribution in football. Specifically, we consider all matches of the clubs participating in the top four European football leagues during the 2015-2016 Season.

Finding and defining groups of occupations that meaningfully belong together is a never-ending endeavour in Sociology - most notably, research using Social Classes is completely based on this idea that occupations can be meaningfully grouped. A core proposition is that individual's social background channels offspring into particular "classes" based on cultural and economic factors. For example, some social backgrounds equip children with values and means to become doctors and lawyers, while other backgrounds make offspring wanting (and able) to become carpenters or mechanics. The grouping of occupations into "classes", however, is usually done on a theoretical basis that some might call ad hoc.

We take a different approach by looking at cross-classification tables of sibling occupations to inductively infer which occupations co-occur more than by chance.
We do this with Swedish register data of ~1M individuals across ~150 occupations using multilayered stochastic blockmodels for valued networks. This allows us to find groupings of occupations that co-occur among individual from the same social background (i.e. siblings). In a second step, we analyse the cultural and economic factors that underlie the demarcation of the different groups of occupations.

Max Weber observed that both material and ideal interests govern human conduct, but left underdeveloped the structural mechanisms through which they operate. This paper addresses that gap by integrating belief structures into Elementary Theory (ET), a formal network approach grounded in the structural modeling of power. While ET effectively explains strategic action in terms of material constraints and benefits, it has not accounted for the influence of ideas and beliefs on actor decision-making.

This paper extends ET to include status structures that rank positions by influence rather than control over resources. Whereas power structures shape action through material exclusion and differential benefit, status structures operate through belief formation. High-status actors influence how lower-status actors interpret their circumstances, which in turn shapes their ranking of available options. These belief structures may align with or diverge from material reality, but either way, they guide strategic action.

The model shows how exploitative conditions can be reframed through influence. For example, when high-status actors lead others to believe they are engaged in prosocial or equitable exchange, lower-status actors may act against their material interests. These belief effects are formally incorporated into ET’s preference system, allowing for a systematic representation of how material and ideal interests interact. The resulting model provides a composite structure that traces how beliefs about motivations and structural positions influence the flow of power and the shape of strategic action.

By incorporating cultural belief structures into a formal model of networked interaction, this paper offers a structural account of how meaning and material interests intersect. It contributes to the study of socio-semantic networks by revealing how actors interpret and respond to structured inequalities, offering new tools for analyzing the relational architecture of both power and culture.

Alzheimer’s disease and related dementias (ADRD) affect over 100 million adults and family care-partners worldwide, posing interconnected challenges of brain change, social identity, and everyday life. Yet dominant medical and social scientific narratives reduce or individualize these predicaments. Loss-centered framings obscure meaningful experiences and non-linear decline, while “clinical journey” metaphors understate hardship and disparities. These accounts obscure how physiology, culture, biography, and history intersect—shaping the work needed to make sense of and navigate dementia.

Drawing on advances in cultural science, this article examines how people living with dementia (PLWD) and family care-partners navigate ADRD relationally. Using semantic networks alongside qualitative analyses of a team-based, multi-sited ethnography—including 300+ field visits, 100+ interviews, and five years of immersion across three U.S. states—we show essential cultural work around experiences eclipsed by reductionist accounts. We find PLWD and caregivers engage in pragmatic sense-making, drawing on and reconfiguring shared meanings to respond to practical dilemmas. Rather than resolving contradictions through an overarching illness narrative, this process enables dynamic adaptation to challenges of disease progression, social relationships, and healthcare encounters.

To aid analysis, we fine-tuned a transformer-based language model (RoBERTa) on 17,726 paragraphs of dementia discussions from transcripts and generated semantic networks to visualize “cultural schemata”—frameworks for making sense of experiences through language. Four shared schemata anchor in-depth narratives: health, cognition, social roles, and change. Case comparisons reveal divergences—from disparate care options to atypical disease progression—which influence how schemata are invoked. Yet, pragmatic sensemaking is found across regions and sociodemographic categories as both adaptation and agency.
We conclude by showing how our findings and computational-qualitative approach—extending mixed-methods principles to connect, triangulate, and interpret multilayered data that move beyond dualism—adding needed precision to studies at the nexus of culture, health, and inequality, with implications for science and practice.