Network Modeling for Empirically-Oriented ABMs
Presentation about approaches to and challenges in generative statistical modeling of network data as an ingredient in Agent-Based Models.
Presentation about approaches to and challenges in generative statistical modeling of network data as an ingredient in Agent-Based Models.
An intensive 5-day workshop on social network analysis with R going through concepts, methods and practice of social network analysis: network data, descriptive analysis and elements of statistical modeling.
The distribution of the number of acquaintances among members of a society is a relevant feature of its social structure. Furthermore, the number of acquaintances (or “degree”) is used for estimating other societal features, such as the size of hard-to-count subpopulations or social cohesion. To estimate the degree, the Network Scale-Up Method (NSUM) asks survey respondents about the number of people they know with a set of first names for which name statistics are available. For this method to be precise, a set of names needs to be selected for the survey that jointly represent the population on a smaller scale in terms of relevant traits such as gender or age. Finding the optimal set of names is a combinatorial problem for which this paper provides a solution approach. The approach can serve other NSUM users, and can be applied to any population for which name statistics distributed over different categories are available. We empirically show that our approach successfully provides subsets of names replicating the population distribution for six countries with very different name statistics.
Theory and practice of fitting Exponential-family Random Graph Models to egocentrically sampled network data. Together with Statnet team.
An introdctory workshop to SNA tools in R. Together with Lorien Jasny.
The distribution of the number of acquaintances among members of a society is a relevant feature of its social structure. Furthermore, the number of acquaintances (or “degree”) is used for estimating other societal features, such as the size of hard-to-count subpopulations or social cohesion. To estimate the degree, the Network Scale-Up Method (NSUM) asks survey respondents about the number of people they know with a set of first names for which name statistics are available. For this method to be precise, a set of names needs to be selected for the survey that jointly represent the population on a smaller scale in terms of relevant traits such as gender or age. Finding the optimal set of names is a combinatorial problem for which this paper provides a solution approach. The approach can serve other NSUM users, and can be applied to any population for which name statistics distributed over different categories are available. We empirically show that our approach successfully provides subsets of names replicating the population distribution for six countries with very different name statistics.
A keynote at ARS'23 with a call to measuring acquaintanship networks and the preliminary results and plans for the PATCHWORK project.
An introdctory workshop to SNA tools in R. Together with Lorien Jasny.
The workshop will talk you through more advanced features of igraph such as node and edge indexing, segregation measurement and community detection.
A poster at EUSN 2022 on estimating egocentric ERGMs with marriage ties as a covariate.
The workshop will talk you through more advanced features of igraph such as node and edge indexing, segregation measurement and community detection.
An introdctory workshop to SNA tools in R. Together with Lorien Jasny.
A Network Science Approach to Social Cohesion in European Societies. An ERC-funded project lead by Miranda Lubbers (UAB).
R package for representing graphs as graph6, digraph6 or sparse6 strings.
Social networks & legal devices.
Mapping the Portuguese Atlantic Empire. Reconstructing and mapping networks of official correspondence of the early modern Atlantic Portuguese Empire.
R package with measures of homophily and segregation in networks.
R package for proximity-based methods of link prediction.
R Package for importing and analysing ego-centered-network data
Statnet is a suite of R packages for the analysis, simulation and visualization of network data focused around Exponential-family Random Graph Models. It is developed by Pavel Krivitsky, Mark S. Handcock, David R. Hunter, Carter T. Butts, Chad Klumb, Steven M. Goodreau, and Martina Morris plus contributors (whom I enjoy beeing).
R package for drawing alluvial diagrams.
Dynamics of Competition and Collaboration in Science: Individual Strategies, Collaboration Networks, and Organizational Hierarchies. Funded through NCN Sonata program. Website archived.
R package for converting network data objects between different classes.
System Analizy Orzeczeń Sądowych. A search engine and analysis platform of judgemenets issued by Polish courts.
Warsaw School of Data Analysis (WSAD). Website archived.
Dynamics of Cooperation, Networks and Institutions (archived).