Computational Social Complexity

Thematic Track at 2023 (webpage)

These notes can be found here:
https://tinyurl.com/CSCx2023notes

We are having the discussion at the hanau pavilion https://maps.google.com/?cid=13656252924571330941&entry=gps

Overview

Humans live in complex societies based on extensive cooperation, connected through social and asocial relationships. These interactions generate numerous emergent patterns. As a theoretical tool, social complexity theory provides a platform for hypothesis testing regarding the emergence of macro- and meso-level social phenomena. The track aims are to stimulate interdisciplinary research to develop complex systems-based approaches aimed at understanding social systems. We will focus on the following questions: What are the patterns in social systems that existing theories and data cannot explain? What kinds of observational and empirical data are needed to better inform the models? What new modeling techniques and methods need to be developed?

Broad topics in social complexity for which we welcome papers:

1.     Modeling for sustainable development goals (e.g., poverty, well-being, food security, water, energy) in rapidly growing urban complex societies, focusing on the role of modeling in policy and urban planning.

2.     Emergent phenomena of complex social systems: cooperation, self-organization, regime shifts, tipping points, and resilience.

3.     Social Digital Twins and The Social Complexity of Digital Data (complex realism in social research, new metatheory), including novel (curated) datasets and empirical calibration and validation of computational models for understanding complex social systems.

4.     New modeling techniques in computational social science.

Presentations:

-In-person session-

Title: Toxicity in Evolving Twitter Topics

Authors: Marcel Geller, Flávio L. Pinheiro and Vítor V. Vasconcelos

Abstract: This paper investigates the relationship between the topic evolution and speech toxicity on Twitter. We construct a dynamic topic evolution model based on a corpus of collected tweets. A combination of traditional static Topic Modelling approaches and sBERT sentences Embeddings are leveraged to build a Topic Evolution Model that is then represented as a directed Graph. Furthermore, we propose a hashtag-based method to validate the consistency of a Topic Evolution Model and provide guidance for the hyperparameter selection. We identify five evolutionary steps -- Topic Stagnation, Topic Merge, Topic Split, Topic Disappearance, and Topic Emergence. Utilizing a speech toxicity classification model, we analyze the dynamics of toxicity in the Evolution of Topics. In particular, we compare the aforementioned Topic Transition Types in terms of their toxicity. Our results indicate a positive correlation between the Popularity of a Topic and its Toxicity. The different transition types do not show any statistically significant difference in the presence of inflammatory speech.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• Provides temporal dynamics of topic toxicity at aggregated level and shows a popularity-toxicity space in which topics evolve.
• The study was performed in a limited dataset of twitter topics collected for New York in 2019-2020.
• This collection was particularly problematic since the geographical location made the tweets more sparse in time than otherwise desirable for studying temporal dynamics.
• In reality Topics can arguably not only emerge, but also re-emerge.
• The Topic Evolution Model presented in this paper could be extended by not only comparing Topics of direct adjacent time-bins, but also previous time-bins. This would remove the Mpartite property of the Graph but enable even more profound understanding of the Topic Evolution.
• No information was used about the topics existing in the previous timestep (time intervals were independent); this could help with identifying continuity of topic evolution

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Better identification of topic trajectory in the toxicity-popularity space can help optimize intervention timing

Unanswered questions and future research:

• We need mechanistic descriptions/theories for large scale topic popularity and its dynamics. These methodologies can provide the data and be co-developed for better model and data alignment.
• Extension to other languages?
• Twitter is known for its tendency to amplify extreme viewpoints. Users often gravitate towards positions that align with their own beliefs, forming echo chambers that reinforce their opinions. This can lead to a lack of productive dialogue and an increase in toxic exchanges between opposing sides.Modelling these processes are important?
• Rapidly evolving topics often lack verified information, leading to the spread of rumors, misinformation, and even deliberate disinformation campaigns. How to regulate this?

Title: Longitudinal Analysis of the Topology of Criminal Networks using a Simple Cost-Benefit Agent-Based Model

Authors: Louis Weyland, Ana Isabel Barros and Koen van der Zwet

Abstract: Recently, efforts have been made in computational criminology to study the dynamics of criminal organisations and improve law enforcement measures. To understand the evolution of a criminal network, current literature uses social network analysis and agent-based modelling as research tools. However, these studies only explain the short-term adaptation of a criminal network with a simplified mechanism for introducing new actors. Moreover, most studies do not consider the spatial factor, i.e. the underlying social network of a criminal network and the social environment in which it is active. This paper presents a computational modelling approach to address this literature gap by combining an agent-based model with an explicit social network to simulate the long-term evolution of a criminal organisation. To analyse the dynamics of a criminal organisation in a population, different social networks were modelled. A comparison of the evolution between the different networks was carried out, including a topological analysis (secrecy, flow of information and size of largest component). This paper demonstrates that the underlying structure of the network does make a difference in its development. In particular, with a preferentially structured population, the prevalence of criminal behaviour is very pronounced, giving the criminal organisation a certain efficiency.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• Combines an agent-based model with an explicit social network to simulate the long-term evolution of a criminal organisation. It addresses shortcomings of previous studies that focus on short-term adaptation or neglect the social network.
• Demonstrates the importance of the underlying network structure in the evolution and efficiency of a criminal organisation.
• Theoretical study, suggesting a need for further empirical research to validate the model against real-world data.
• Assumes a static network weight throughout the simulation, which may not reflect the dynamic nature of social ties and their influence over time.
• There is a gray zone in criminology -- not honest vs criminal

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Findings could be valuable for law enforcement and policymakers as they offer a fresh perspective on the development of criminal organisations over the long term. This could support the development of more effective strategies for combating criminal networks.
• Emphasizes the importance of considering the social context and structure when modeling the evolution of criminal behaviour within a population, which can have practical implications for how interventions are designed and implemented.

Unanswered questions and future research:

• The study raises questions about the ideal types of observational and empirical data that would best inform this and similar models. The authors suggest that longitudinal data on criminal networks, combined with social network data collected by a national statistical office, would be ideal.
• The current study does not consider changes in the strength of relationships between actors over time, which could be an area for future research. The introduction of dynamic changing weights in the model could offer a more accurate representation of the evolving social ties within criminal networks.
• The authors also indicate that local clustering could play a significant role in the spreading of criminal networks. Future research could therefore examine this claim in more depth, for instance by conducting an analysis of different local communities to understand how criminal behaviour spreads within them.
• More broadly, the study suggests the need for the development of new modeling techniques that can better account for the long-term evolution of criminal networks and the role of social networks in this process. Future research could focus on these aspects to enhance the existing body of knowledge in computational criminology.
• With more investigation and refinement, agents’ honesty oscillations may help mirror the “grey zone” behavior observed in real criminal networks.

  ---

Title: Manifold Analysis for High-Dimensional Socio-Environmental Surveys

Authors: Charles Dupont and Debraj Roy

Abstract: Recent studies on anthropogenic climate change demonstrate a disproportionate effect on agriculture in the Global South and North. Questionnaires have become a common tool to capture the impact of climatic shocks on household agricultural income and consequently, on farmers’ adaptation strategies. These questionnaires are high-dimensional and contain data on several aspects of an individual (household) such as spatial and demographic characteristics, socio-economic conditions, farming practices, adaptation choices, and constraints. The extraction of insights from these high-dimensional datasets is far from trivial. Standard tools such as Principal Component Analysis, Factor Analysis, and Regression models are routinely used in such analysis. However, the above methods either rely on a pairwise correlation matrix, assume specific (conditional) probability distributions in its construction, or assume that the high-dimensional survey data lies in a linear subspace. Recent advances in manifold learning techniques have demonstrated better detection of different behavioural regimes from surveys. This paper uses Bangladesh Climate Change Adaptation Survey data to compare three non-linear manifold techniques: Fisher Information Non-Parametric Embedding (FINE), Diffusion Maps and t-SNE. Using a simulation framework, we show that FINE appears to consistently outperform the other two methods except for questionnaires with high multi-partite information. While not being limited by the need to impose a grouping scheme on data, t-SNE and Diffusion Maps require some tuning and thus more computational effort since they are sensitive to the choice of hyperparameters, unlike FINE which is non-parametric. Finally, we show that FINE is able to detect adaptation regimes and corresponding key drivers from high-dimensional data.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• The study performs sensitivity analysis across methods 2D embedding of community survey data
• Sensitivity to hyperparameter tuning dependent on the dimensionality of the underlying truth
• Discretization of continuous variables -- method uses discrete data

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• FINE assists with visualisation and information/insight extraction from high dimensional survey data which can be applied across many fields (e.g., sociology, medicine, economics, etc.).

Unanswered questions and future research:

• Different types of non-linearities?
• Iteration process between the output of this analysis and

• Are there lessons we can learn for (re-)designing questionnaires from the analysis results?
• What types of data does the model require?
• Can this be used to benchmark surveys ahead of their deployment?

  ---

Keynote: Building Agent-Based Models for policy support based on Qualitative Inquiry. The case of Disaster Information Management in Jakarta, Indonesia

Keynote: Vittorio Nespeca, University of Amsterdam & TU Delft

Abstract: Qualitative research is a powerful means to capture human interactions and behavior. Although there are different methodologies to develop models based on qualitative research, a methodology is missing that enables to strike a balance between the comparability across cases provided by methodologies that rely on a common and context-independent framework and the flexibility to study any policy problem provided by methodologies that focus on capturing a case study without relying on a common framework. In this keynote, I will present a methodology targeting this gap for ABMs in two stages. First, a novel conceptual framework centered on a particular policy problem is developed based on existing theories and qualitative insights from one or more case studies. Second, empirical or theoretical ABMs are developed based on the conceptual framework and generic models. This methodology is illustrated by an example application for disaster information management in Jakarta, resulting in an empirical descriptive agent-based model.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• Balance cross-case comparability and flexibility to capture different policy problems and agent architectures, across case studies.

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Policies needed to foster the effective exchange of information in crises
• Both formal nomination of info exchange hubs and increasing density of intergroup social ties through community engagement are effective interventions to foster the effective exchange of information

Unanswered questions and future research:

• Mechanisms for the emergence of intergroup info exchange hubs
• Policies that foster such an emergence to improve info. Quality
• Is there a role for an iterative process? Redesign the framework, recalibrate models, redefine communities…
• Can crisis management improve regular work of organizations
• Exercises with help services/organizations would provide additional and useful data, but that is time-consuming and costly -- how does the model learn from those, then?

-Online session-

Title: The social graph based on real data

Authors: Tomasz Gwizdałła and Aleksandra Piecuch

Abstract: In this paper, we propose a model enabling the creation of a social graph corresponding to real society. The procedure uses data describing the real social relations in the community, like marital status or number of kids. Results show the power-law behavior of the distribution of links and, typical for small worlds, the independence of the clustering coefficient on the size of the graph.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• 4 levels of social connections:

• household members,
• close connections outside household,
• acquaintances,
• random/possible contacts

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Not discussed

Unanswered questions and future research:

• Other types of interactions (e.g., higher education) are eventually to be addressed.

Title: The process of polarisation as a loss of dimensionality.

Authors: Sage Anastasi and Giulio Valentino Dalla Riva

Abstract: The increasing polarisation in our society, and in particular in our Social Networks, has been the focus of much research, especially during the Sars-Cov-2 pandemic. Polarisation is widely believed to be a risk for our democracies. Understanding and detecting its temporal evolution is, therefore, highly important.

Current approaches to define and detect polarisation largely rely on finding evidence of bimodality in a (possibly latent) ideological distribution, often inferred through collective behaviours on Social Networks. Bimodality-based definition makes it hard to detect temporal trends in polarization, as the relevant tests deliver results that fall into a binary of polarised or non-polarised. Building on post-structuralist understanding of polarisation processes, we propose here an alternative definition and estimate technique for polarisation: it is a decrease in the dimensionality of the latent space underpinning a communication network. This allows us to have a more nuanced definition of polarisation, apt to detect its increase or decrease beyond a binary detection test.

In particular, we exploit the statistical theory of Random Dot Product Graphs to embed networks in metric spaces. A decrease in the optimal dimensionality for the embedding of the network graph, as measured using truncated singular value decomposition of the graph adjacency matrix, is indicative of increasing polarisation in the network. We apply our framework to the communication interactions among New Zealand Twitter users discussing about climate change, from 2017. In line with our expectations, we find that the discussion is becoming more polarised over time, as shown by a loss in the dimensionality of the communication network and corroborated by a decrease of the Von Neumann complexity of the network. Second, we apply this analysis to discussions of the COP climate change conferences, showing that our methods agree with other researchers' detections of polarisation in this space. At the end, we provide some synthetic examples to demonstrate how an increase of the isolation between distinct communities, or the increase of the predominance of a community other the others, in the communication networks are identifiable as polarisation processes.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• Create a social graph from social media (but Twitter and Reddit are now limiting access to its data. )
• The underlying mechanism is not clear -- are mixed group less polarized or do they polarize less because they are mixed? How is a group without connections polarized if they don’t interact with others?
• Interactions are always assumed to be positive

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Not discussed

Unanswered questions and future research:

• Do models of network formation not reproduce the patterns you find in the data? What types of mechanisms do they consider?
• Signed matrices for SVD. Do they provide similar information? More useful information?

Title: Structural Validation Of Synthetic Power Distribution Networks Using The Multiscale Flat Norm

Authors: Rounak Meyur, Lyman Kostiantyn, Bala Krishnamoorthy and Mahantesh Halappananvar

Abstract: We study the problem of comparing a pair of geometric networks that may not be similarly defined, i.e., when they do not have one-to-one correspondences between their nodes and edges. Our motivating application is to compare power distribution networks of a region. Due to the lack of openly available power network datasets, researchers synthesize realistic networks resembling their actual counterparts. But the synthetic digital twins may vary significantly from one another and from actual networks due to varying underlying assumptions and approaches. Hence the user wants to evaluate the quality of networks in terms of their structural similarity to actual power networks. But the lack of correspondence between the networks renders most standard approaches, e.g., subgraph isomorphism and edit distance, unsuitable. % for this purpose. We propose an approach based on the multiscale flat norm, a notion of distance between objects defined in the field of geometric measure theory, to compute the distance between a pair of planar geometric networks. Using a triangulation of the domain containing the input networks, the flat norm distance between two networks at a given scale can be computed by solving a linear program. In addition, this computation automatically identifies the 2D regions (patches) that capture where the two networks are different. We demonstrate our approach on a set of actual power networks from a county in the USA. Our approach can be extended to validate synthetic networks created for multiple infrastructures such as transportation, communication, water, and gas networks.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• Hausdorff distance is the metric that has mostly been used for structural comparison; but it has several challenges which have been addressed by the flat norm metric.

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Provides a method to compare a pair of geometries on a plane surface. It can be used to distinguish between pair of networks.
• Synthetic electric networks are important to policymakers for planning purposes; but they need to know which synthetic networks resemble actual networks more accurately.

Unanswered questions and future research:

• At least one accurate network is required to measure the resemblance - but this is an unanswered problem for any sort of validation.

• Can we use methods for sensitive datasets to generate social networks too? For powernetworks we need to incorporate space, for some types of social connections these should be important too -- similar problem to covert networks

Title: OptICS-EV: A Data-Driven Model for Optimal Installation of Charging Stations for Electric Vehicles

Authors: Kazi Ashik Islam, Rounak Meyur, Aparna Kishore, Swapna Thorve, Da Qi Chen and Madhav Marathe

Abstract: As the demand for electric vehicles continues to surge worldwide, it becomes increasingly imperative for the government to plan and anticipate its practical impact on society. In particular, any city/state needs to guarantee sufficient and proper placement of charging stations to service all current/future electric vehicle adopters. Furthermore, it needs to consider the inevitable additional strain these charging stations put on the existing power grid. In this paper, we use data-driven models to address these issues by providing an algorithm that finds optimal placement and connections of electric vehicle charging stations in the state of Virginia. Specifically, we found it suffices to build 10,733 additional charging stations to cover 75% of the population within 0.33 miles (and everyone within 5 miles). We also show optimally connecting the stations to the power grid significantly improves the stability of the network. Additionally, we study 1) the trade-off between the average distance a driver needs to travel to their nearest charging station versus the number of stations to build, and 2) the impact on the grid under various adoption rates.These studies provide further insight into various tools policymakers can use to prepare for the evolving future.

Strengths and weaknesses of the study in relation to the relevant literature, with a particular focus on identified gaps and mismatches in results:

• The optimal cover solution exists for a specific number of stations. Or for a specific radius. But they will be built sequentially. At what timescale should we consider optimility, then?
• How do we account for resilience? And which parts of the system are more resilient? (i.e., when a node fails, who loses access? At which distance is the backup? Also, are all charging stations connected to the same power supply (i.e., if one fails, do they all fail/could they cause the power network as a whole to fail, or does the supply network have some level of modularity)?) Is there a fairness issue there too?

Meaning of the study (possible mechanisms and implications for policymakers/clinicians):

• Optimization process. To be inferred from text.

Unanswered questions and future research:

• Heterogeneity/fairness of access as stations are deployed. How can we use these optimization processes in a fair way?

Discussion Points & Relevant Literature

Twitter just shut down the chapter on academic research (Justine Calma, The Verge)

What are the boundaries of Computational Social Complexity?

What are the Challenges of Computational Social Complexity?

Recent advances in ML and Deep Learning tools have shown that complicated (linear and non-linear) are reachable and can be solved with the right combination of Computational Power and Historical Data. The next barrier are complex problems, or problems typically associated with Complexity and/or of Complex Systems. In that case, the current available tools, altough extremely powerful, are still not quite there. This track can be about exploring such boundaries of how we should study Complexity in the context of Social Systems.

Mapping Systems Complexity and Expert Input:

• Given multi-domain multi-causation and heterogeneity, how to define system boundaries?

• Cilliers, P. (2001). Boundaries, hierarchies and networks in complex systems. International Journal of Innovation Management, 5 (02), 135-147.

• How to not reinvent the wheel and make sure we collect and capture domain expert knowledge?

• Barbrook-Johnson, P., & Penn, A. S. (2022). Systems Mapping: How to build and use causal models of systems.

Hidden (multilayered) networks and data

• The observed network may not be relevant for the behavior of interest.
• The relevant network is not directly accessible and must be inferred and so do the types of interactions between agents

• Peel, L., Peixoto, T. P., & De Domenico, M. (2022). Statistical inference links data and theory in network science. Nature Communications, 13(1), 1-15.
• Rossi, E., Monti, F., Leng, Y., Bronstein, M., & Dong, X. (2022, June). Learning to Infer Structures of Network Games. In International Conference on Machine Learning (pp. 18809-18827). PMLR.

• Decision-makers may not have access to network data for privacy reasons; how does inference circumvent or trick the policy?

 Predictability

• "Researchers must reconcile the idea that they understand life trajectories with the fact that none of the predictions were very accurate"

• Salganik, M. J., Lundberg, I., Kindel, A. T., Ahearn, C. E., Al-Ghoneim, K., Almaatouq, A., ... & McLanahan, S. (2020). Measuring the predictability of life outcomes with a scientific mass collaboration. Proceedings of the National Academy of Sciences, 117(15), 8398-8403.

Interventions

• We want to move beyond understanding complex social systems towards acting on them.

• Moore, G. F., Evans, R. E., Hawkins, J., Littlecott, H., Melendez-Torres, G. J., Bonell, C., & Murphy, S. (2019). From complex social interventions to interventions in complex social systems: future directions and unresolved questions for intervention development and evaluation. Evaluation, 25(1), 23-45.
• Valente, T. W. (2012). Network interventions. science, 337(6090), 49-53

Policy agenda

• Cost-Effectiveness/Affordability and acceptability of intervention are tested locally. Is it scalable? Or is it tailored locally? If you target the whole population the intervention is simple. With local communities, the complexities are important.

• Castellani, B., Bartington, S., Wistow, J., Heckels, N., Ellison, A., Van Tongeren, M., ... & Reis, S. (2022). Mitigating the impact of air pollution on dementia and brain health: Setting the policy agenda. Environmental Research, 114362.

Other

• Lazer, D. M., Pentland, A., Watts, D. J., Aral, S., Athey, S., Contractor, N., ... & Wagner, C. (2020). Computational social science: Obstacles and opportunities. Science, 369(6507), 1060-1062. [Abstract: The field of computational social science (CSS) has exploded in prominence over the past decade, with thousands of papers published using observational data, experimental designs, and large-scale simulations that were once unfeasible or unavailable to researchers. These studies have greatly improved our understanding of important phenomena, ranging from social inequality to the spread of infectious diseases. The institutions supporting CSS in the academy have also grown substantially, as evidenced by the proliferation of conferences, workshops, and summer schools across the globe, across disciplines, and across sources of data. But the field has also fallen short in important ways. Many institutional structures around the field—including research ethics, pedagogy, and data infrastructure—are still nascent. We suggest opportunities to address these issues, especially in improving the alignment between the organization of the 20th-century university and the intellectual requirements of the field.]
• Data can come from social media, profiling consumer preferences, or financial status. Phone calls, social media, and field studies have dropping rates of participation (argument by Chris Bail, June 2023 in a workshop at IAS UvA). Will data come from online environments exclusively?

Boundaries and Challenges of Computational Social Complexity

        Non traditional. Complexity is more than puting together two fields. It needs different perspectives. Micro-interactions. Complexity is a big bag that approaches complex systems. Hyphenated. Which subset of issues are relevant. Policymaking in complex systems.

Important distinction: civil engineering using complex systems methods.

Would we define it more clearly? Social actors.

Tools to apply to understand. Deductive reasoning and reductivism.

Define the systems and define the tools. SIR is not. But SIR with policy making.

Operations research is further from this topic.

Social science with complexity methods.

Empirical collection of data. Computational social science --- it is an issue there.

Any study which give mechanistic understanding of a social process. Plus methods to analyse. Plus method to generate dynamical data.

• Given the recent developments in data accessibility like Twitter shutting down academic research, what boundaries do you foresee for Computational Social Complexity?

Not a problem for complexity. We can use mixed methods. Hard to confirm, but a great to narrow down.

• What are the challenges that researchers might face in Computational Social Complexity in terms of data access, model building, or policy implications?

Model credibility. Model-based decisions. Building consensus in the formulationof policies. THen these participatory modeling techniques go in the direction. Convincing (all) people.

• Given the multi-domain, multi-causal, and heterogeneous nature of social systems, how can researchers effectively define system boundaries?

• How can researchers effectively incorporate domain expert knowledge to prevent 'reinventing the wheel'?

Hidden (Multilayered) Networks and Data

• What are some challenges and strategies for dealing with hidden networks and data layers that are relevant but not directly accessible?

Only network science is for another track. Network topology metrics is out. Hidden structures and interactions, but not just hidden structures. Methods for modeling accessing strategic interection.

• What are the implications for data privacy in the context of network data, and how might inference circumvent or trick these policy constraints?

Is there and ethics involved? Can an open model be accessible for unethical things. How to anonimise a model -- abstract information so it is not misused. Example of los-alamos. Possible discussion for 2024.

Predictability and Interventions

• How do researchers reconcile the unpredictability of life outcomes with their models and predictions?

---deep--- Important for interventions.

• Considering that the ultimate goal is not just to understand but to act on complex social systems, what are the prospects for designing effective interventions?

The process of building a model IS about understanding rather than predicting. The tolerance is . Law of parcimony.

Does it help making the decision? Is the participatory modeling more helpful? You can measure learning. Connect to field experiments because of the control.

• Can models help with challenges that arise when trying to scale locally tested interventions, considering the complexities of different local communities?

Policy Agenda

• How can research in Computational Social Complexity inform policy agendas, particularly considering issues like cost-effectiveness, affordability, and acceptability of interventions?

• How can these studies balance between scalability and local tailoring of interventions?

Data Sources

• As participation rates in traditional data collection methods like phone calls, social media, and field studies are dropping, what alternative data sources can researchers rely on?

Maybe data collected locally can be more useful cause they have better context. How do we make those more representative. --- We can use models. If the models considers a process that is sensitive to the network, then collect data on that network.

• Will online environments become the exclusive source of data, and if so, what implications might this have for Computational Social Complexity?
  

Future Directions

• Considering these challenges and questions, what are the potential future directions for research in Computational Social Complexity?
• How could these issues shape the agenda for future work in this field, and what advancements would we like to see?

Target keynote --- even if  they don’t show up

Work on the wording for discussion requisition (it was intimidating)

We should target specific people/invite them.

Start earlier

Engage more people at the UvA.

More people from ecology, and conventional social science, behavioral science (including other animals?), sociology.