Complex Systems Summer School 2017-Projects & Working Groups

From Santa Fe Institute Events Wiki

Complex Systems Summer School 2017

Note: if you are not going to continue one of the project ideas below during the CSSS please put them in the archived projects section below.


Evolution Working Group/Microbial Systems are Complex

Meeting on Thursday 6/15 at 7:30


Several of us seem to have shared interests in microbial ecology/evolution/genetics and/or evolutionary ecology/evolutionary theory in general. It might be a good idea to meet up and brainstorm a project (maybe over beer?). Some potential ideas that were discussed during speed dating:

  • The evolution of the bacterial pangenome
  • Development/adaptation of classical population genetic techniques to microbes (maybe via a network describing HGT interactions)
  • The evolution of metabolism/the interaction between metabolism and community structure
  • How interspecific interactions (e.g. crossfeeding, competition) influence the rate of decay of community similarity in a homogeneous environment (scale of spatial organization)

Interested Participants

IF INTERESTED PLEASE JOIN SLACK GROUP evolution_microbes (Will not be emailing people on this list as it grows)

  • Jake Weissman (jw4336(at)
  • Makoto Jones (makoto.jones(at)
  • Hilje Doekes (h.m.doekes(at)
  • Alicia Kraay (amullis (at)
  • Ximo Pechuan
  • Jiří Moravec
  • Yael Gurevich (yael01[at]


EU: responding to migrant crisis

Next meeting & Homework

Next meeting: Monday at lunch. Meet up in Great Hall after lecture and head to lunch together


  • familiarize yourself with NetLogo model "Diffusion on a directed network" (in NetLogo library)
  • come up with a list of factors (e.g. GDP, geo...) that you think most important/worth considering
  • read Scheffer et al (see "Literature") & 1 other paper of your choice (will quickly brief the group in it on monday)


Keywords: cooperation, migration, decision-making consensus-reaching, supranational unions

Research question: In the context of the 2015 European migrant crisis, why did it take so long for the EU to reach an agreement on how to deal with the migrant crisis? What factors might have determined this delay and general lack of efficiency? Potential factors that might be considered: geography, GDP, migrant pressure, cost of welcoming migrants... (will be determined during Monday meeting)

Methods: adapt NetLogo model "Diffusion on a directed network" to address the question above. First phase: explore how the system works with arbitrary parameters. Second phase: plug in real data


Potential data

Interested Participants

  • Alberto Micheletti (biology, social evolution, ajcm2[at]st-andrews[dot]ac[dot]uk) - group contact
  • Andrew Johnson (biology, ecology, competition, collapse, afjohnson[at]ucsd[dot]edu)
  • Ella Jamsin (sustainability, design, tipping points, e[dot]jamsin[at]tudelft[dot]nl)
  • Basak Taraktas (
  • Hilje Doekes (theoretical biology, evolution, microbial cooperation and spite, h.m.doekes (at)
  • Laura Elsler
  • Maartje Oostdijk
  • Martina Balestra (management, human computer interaction, mb5758[at]nyu[dot]edu)
  • Madison Hart (psychology, social evolution, collapse, madisonkhart(at)

Gut-Brain Axis: Impact of gut microbiota on mental health

BRAINSTORMING MEETING OVER DINNER on 6/15 (THURSDAY=TODAY). MEET IN CAFETERIA. Email Bleu (happylittleaccident AT gmail) with questions


public health, microbiome, metabolism, network analysis, neuroscience


The term "gut feeling" is not without a scientific basis. Recent literature has emphasized the connection between the brain and intestinal microbes. Studies are beginning to link neurodegenerative disorders, such as Alzheimer's disease and multiple sclerosis, to gut dysbioses.

Other relationships between the brain and the gut are waiting to be explored. Also, on a broader level, relationships between microbial metabolism and neurotransmitter levels could be investigated.

Open source data sets are available, such as the ones mentioned in this paper:

If you find this interesting, include your name and email on the list below

Interested Participants

  • Bleu Knight (bleu AT; happylittleaccident AT
  • Aaron Schwartz (aschwa15 AT
  • Madison Hart (madisonkhart(at)
  • Deepa Rao []
  • Rachel Gicquelais (rgic (at)
  • Kyle Reing (

Beer: evolution through the ages


beer, evolution, genetic elements, history


Brewing beer is an ancient art that has experienced a renaissance in the last few decades. Can we model beer recipes in an "evolutionary" sense? Can ingredients like hops and malt be modeled like genes that are expressed more in some beers than in others? Can we create phylogenetic trees that delineate relationships between beers? Can beers be considered 'species', and if so, what has contributed to the diversity that we see in the US today?

If you find this interesting, include your name and email on the list below


beer recipe database:
medieval english ales:
about food (not beer): Flavor network and the principles of food pairing DP: I think it is a great reference to start with!

Interested Participants

Bleu Knight (bleu AT; happylittleaccident AT

  • Doheum Park (park154 AT, innovation, network science, data science)
  • Zhiya Zuo (zhiya-zuo[at]uiowa[dot]edu)

International Trade Patterns in Local Fisheries

Keywords: cascading effects, overfishing, international seafood trade, exploitation patterns


Seafood is an increasingly globalized food commodity with now 40% traded internationally. At the same time, the majority of global fish stocks are being fully- (58.1%) or over-exploited (31.4%). International buyers connect to local fisheries as the fishery is overexploited the buyer moves on to the next location. This pattern of serial exploitation has been observed in a fisheries worldwide such as tuna, cod, sea cucumber. We aim to explore the linkages and changes in bilateral trade networks that happen before and after a fish stock collapse to find different patterns of exploitation. A starting point for the analysis may be studying anomalies and changes in the global network of seafood trade.

Include your name and email on the list below to talk more :)

Meeting Lunch Thursday June 15. Anybody interested in brainstorming, join!

Resources UN Comtrade publishes freely available bilateral trade data:

RAM legacy is a database which contains marine stock assessments (collapses):


Maartje Oostdijk (group contact,
Laura Elsler (group contact)
Basak Taraktas (
Junfu Zhao (
Elaine Bochniewicz (
Ximo Pechuan
Anand Nair (
Greg Britten (part-time,
Alberto Micheletti (ajcm2[at]st-andrews[dot]ac[dot]uk)

Agent-Based Prediction Competition

Report of the meeting of Wednesday 14/06:

Next meeting: Thursday 15/06 over lunch or at 16.15 (to be decided)


Agent-Based Models, Machine Learning, Prediction


Agent-Based Models (ABMs) have so far been used mainly for descriptive purposes. Some people (including myself) think that ABMs can unleash their full potential in making predictions. I envision this project as an open-ended test to this hypothesis. The main idea is borrowed from the machine learning prediction competitions (, in which you are given a training dataset and have to outperform the other participants in predicting the outputs of a test set. Instead of using a machine learning algorithm to approximate a target function, I propose we use an ABM. This project can take several directions:

  • The topic of the ABM and the data can come from any discipline: biology, epidemiology, economics, sociology, ecology, etc.
  • The data can be real or synthetic. In the latter case, someone could code a very complicated ABM as the Data Generating Process of a fictitious system, and the task would be to approximate this complicated ABM.
  • A humble goal would be to take machine learning as an upper bound for prediction possibilities (, and see how putting realistic assumptions in the ABM makes the prediction score close to the machine learning benchmark.
  • A more ambitious goal would be to beat the machine learning algorithms. We know they can fail because the world is non-stationary (see Google Flu Trends failure). ABMs can account for that.
  • Any other direction!

I would be happy to discuss about any practical example with whoever is interested. I think this project can be very risky, but this is why I’m proposing it here! Still, at least a proof of principle would be a great result. I'm open to any hijacking of this project however!

Group Contact

Marco Pangallo (complexity economics, marco[dot]pangallo[at]maths[dot]ox[dot]ac[dot]uk)

Interested Participants

  • Meghan Galiardi (meghangaliardi(at)
  • Madison Hart (madisonkhart(at)
  • Martina Balestra (mb5758[at]nyu[dot]edu)
  • Uzay Cetin (uzay00(at)
  • Emily S (eshouppe [at] gmail [dot] com)
  • Cigdem Yalcin (cigdem_yalcin(at)
  • Chris Miles (
  • Andrew Johnson (afjohnson[at]
  • Stefan B (sfb311[at]nyu[dot]edu)
  • Yuji Saikai (saikai(at)
  • Yael Gurevich (yael01[at]
  • Katarina Mayer (

AI-aided Graph Auto-generation


Graph Auto-generation; Graph Characteristics; Artificial Intelligence; Combinatorial Optimization;


Complex networks/graphs appear in almost every aspect of science and technology, ranging from social systems, computer networks, biological networks to the state spaces of physical systems, even Bayesian Belief Networks. Graph auto-generation has wide applications. Just like what is done by the Google AlphaGo, this oftentimes ends up as a combinatorial optimization problem. The prohibitive number of possibilities calls upon the Artificial Intelligence to determine better strategic choices at every turn. There are several interesting topics.

  • As a starter, we would like to determine the topological structure of given networked systems. A variety of basic measures and metrics are available that can tell us about small-scale structure in networks, such as correlations, connections and recurrent patterns, but it is considerably more difficult to quantify structure on medium and large scales which can help us understand the ‘big picture’.
  • Next, a fundamental property of interdependent networks is that failure of nodes in one network may lead to failure of dependent nodes in other networks. This may happen recursively and can lead to a cascade of failures. In fact, a failure of a very small fraction of nodes in one network may lead to the complete fragmentation of a system of several interdependent networks.
  • Thirdly, assume we have a collection of node and a defined dependency model. Can we auto-assemble these node to form a network which satisfies performance requirements. The dependency model can be defined in the following way
   1. Each node has a defined interface, which specifies what kind of supplies it needs and what kind output it can produce.
   2. The output of a given node can be computed based on its chosen suppliers and available supply levels.
  • Finally, given a set of nodes, assuming the edge length between any pair of nodes is computable, can we find a shortest path between two given nodes such that the path consists of edges with bounded length. Moreover, given multiple measures of the edge length, can we find a Pareto front of shortest paths? In other words, a multi-objective optimization of such paths.

Group Contact

Huang Tang (MITRE, Co.)

Interested Participants

  • Sheri Xue Guo (, computer science)
  • Junfu Zhao (

Shared rituals, memes and group identity


meme, rituals, group identity


Shared language or rituals are important for identification of group members from non-members, which itself is important for group cohesion and group identity. In pre-state societies, where survival was dependent on intragroup collaboration, this was of high importance. In modern culture, mere survival is often guaranteed by states, but group identity is still important, whether in meatspace or in cyberspace.

This project will try to identify origin and spread of internet memes, as a form of internet rituals, their importance for group cohesion and their dead, i.e., when popular memes spread into different communities and became common in internet, while losing their original function.

Group Contact

Jiří Moravec j.moravec [at]

Join #memes channel on slack


I think this website is going to be of great help:



Interested Participants

  • shanee stopnitzky - i was going to propose the exact same project! awesome!
  • Rémi Lamarque -

Multi-Layer Networks


Hey all, here is the link to the google slides for Friday 6.16.17 [1]!

  • Link to multi-layer network Dropbox folder: [2]!
  • Link to the full published multi-layer networks in ecology [3]!
  • Link to Sahara-distance-pollinator dataset: [4]!


networks, ecology, multilayer networks, interactions


Networks are a powerful approach to studying ecological systems; however, they typically represent a "monolayer network" of interactions in a snapshot of time. This project is interested in using a "multilayer network" approach to include more complexity in ecological networks. With a multilayer network, we can represent the same ecosystem using layers that describe the system either with several interaction types (competition, cooperation, predation, etc.), communities of species, or over time (time-series data on occurrences, for example). By using interlayer edges, we can connect the monolayers of the network.

The paper that describes this approach includes a table with ~10 available datasets (Supplementary Table 1) and leads to packages in Python, R, and Matlab (Supplementary Table 3).

Some questions include:

  • Using temporal variation in the networks -- what are the dynamics of network assembly?
  • How do different interaction types affect network stability?
  • How do interactions between food webs (via common species) affect the dynamics of each food webs separately?
  • How do processes at one level affect those in other levels?
  • How do diseases spread in interconnected populations of different hosts?

Resources: [5]!


  • The structure and dynamics of multilayer networks-Bocalletti et al. 2014
  • Graph Product Multilayer Networks: Spectral Properties and Applications- Sayama 2017
  • Community Detection in Temporal Multilayer Networks, with an Application to Correlation Networks-Bazzi et al. 2015
  • Disease Localization in Multilayer Networks - Ghillerme et al. 2017
  • The multilayer nature of ecological networks- Pilosof Nature Ecology and Evolution.
  • Kéfi , S., Miele, V., Wieters, E. A., Navarrete, S. A. & Berlow, E. L. How structured is the entangled bank? e surprisingly simple organization of multiplex ecological networks leads to increased persistence and resilience. PLoS Biol. 14, e1002527 (2016). This is study illustrates that trophic and non-trophic interactions in a ‘node-aligned’ multiplex food web are non-randomly organized and that this organization can have important consequences for the persistence of species in a community.
  • Kivelä, M. et al. Multilayer networks. J. Complex Networks 203–271 (2014). This review provides a detailed description of multilayer networks, relevant diagnostics, and models. It gives a starting point to learn about multilayer networks.
  • Boccaletti, S. et al. e structure and dynamics of multilayer networks. Phys. Rep. 544, 1–122 (2014). This review, which takes a di erent perspective from ref. 14, is another starting point to learn about multilayer networks.
  • Gilarranz, L. J., Sabatino, M., Aizen, M. & Bascompte, J. Hot spots of mutualistic networks. J. Anim. Ecol. 84, 407–413 (2014). This study illustrates that the structure of local communities in a metacommunity represented as a multilevel network is a ected by the structure of the network.

Interested Participants

  • Deepa []
  • Basak Taraktas (
  • Bleu Knight [happylittleaccident AT gmail]
  • Shing Zhan []
  • Surendra Hazarie []
  • Martina Balestra [mb5758[at]nyu[dot]edu]
  • Anand Nair []
  • Ximo Pechuan []
  • Kayla Sale []
  • Huang Tang
  • Horacio Marchand (
  • Valentina Baccetti []

Crook's hyperensembles to improve Maximum Entropy Theory of Ecology (METE) and Max Ent in other non-equilibrium systems


METE is a promising approach to detecting macroecological patterns like species abundance and size from relatively small amounts of data, but fails to accurately predict interactive phenomena like size-density relationships. The non-equilibrium nature of open ecological systems is a likely constraint for using max ent for ecological applications. These may be better resolved using Crook's hyperensemble method, which is basically taking the maximum entropy of the maximum entropy of the distribution.

This project is simple, can be applied to many different types of 'ecological systems', data is available and abundant, and we should be able to reach a publishable conclusion during the program.

Interested Participants

  • Shanee Stopnitzky
  • Marco Pangallo
  • Alje van Dam
  • Tim Kunisky
  • Elliot Nelson
  • Chris Miles
  • Deepa Rao
  • Ximo Pechuan
  • Markus Junginger
  • Kyle Reing

Cryptocurrency market predictions


Can we accurately predict cryptocurrency markets using nonlinear time series analysis? Does social network 'popularity' improve predictions?

Interested Participants

  • Shanee Stopnitzky
  • Abdel Abdelgabar (yomegaland AT gmail)

Complex phenomena in cetacean (whale and dolphin) communication


Click sounds produced by whales and dolphins for echolocation and communication are highly complex and possess a fractal structure. Are there discernible patterns in these sounds that could help us decode how these animals perceive the world, communicate with each other and other species (like us?) This needs more thought, but lots of sound data exists.

Interested Participants

  • Shanee Stopnitzky
  • Elaine Bochniewicz
  • Kayla Sale

From Dictatorship to Anarchism (and everything in between)


Politics, social networks, agent based modeling, network theory, network analysis


Study different types of socio-political organizations through network theory and agent based models.
Potential questions: - What kind of structures emerge as a function of, size, agents characteristics such as wealth/power distribution, etc.?
- How does different social structures contribute affect the output of different quantities such as production, self-fulfillment, ... ?
- How does the system evolve when a leader is taken down?
- Does leadership emerge?
- ...

If you find this interesting, include your name and email on the list below

To Do

By Friday, June 15

- Think about how to characterize our agents and the interactions.
- Literature review: every person reads 1 paper/topic that will be presented on Friday.
- Think about what political regimes we want to study and how to describe them.
- Is there data we could use to inform our model-building or to validate our simulation results?



 MASON (Java) -- 
RNetLogo (R) --
Mesa (Python) --
Swarm (Java and Objective C )--
Agentscript --


  1. Standard Cross Cultural Sample (SCCS) -- large ethnographic datasets:
  1. Binford's hunter-gather database:


  1. (Junfu)Veblen, T. (1934). “The Beginnings of Ownership” in his Essays in Our Changing Times.
  2. (Junfu)Field, B. (1989). “The Evolution of Property Rights,” Kyklos 42: 319-45.
  3. (Jose) Foley, D. - Socialist alternatives: from Vienna to Santa Fe
  1. (Junfu)Epstein and Axell, "Growing Artificial Societies- Social Science from the Bottom Up"
  2. (Hugo) Epstein, J. - "Why Model?", Journal of Artificial Societies and Social Simulation vol. 11, no. 4 12
  3. (Hugo) Bartha, R. , Meyerb, M. and Spitznerc, J. , Typical Pitfalls of Simulation Modeling - Lessons Learned from Armed Forces and Business
  4. (Valérie) Epstein J.M.(2002) "Modeling civil violence- an agent-based computational approach". PNAS
  5. (Adrián) Railsback and Grimm, "Agent-Based and Individual-Based Modeling", Chapter 1.
  6. (Damon) Katarzyna Sznajd-Weron and J´ozef Sznajd, "Opinion evolution in closed community"
  7. (Marjan) A Data-Driven Approach for Agent-Based Modeling: Simulating the Dynamics of Family Formation
  8. (Marjan) Agent-Based Modelling Approach for Multidimensional Opinion Polarization in Collective Behaviour
  9. (Marjan) A Model for Revolution Based on Multiplex Networks [CSSS14]
  1. (Marjan) Watts, D. and Dodds, P. - Influentials, Networks, and Public Opinion Formation, (2007)
  1. Marten Scheffer, Steve Carpenter, Jonathan A. Foley, Carl Folke & Brian Walkerk (2001) "Catastrophic shifts in ecosystems", Nature

Interested Participants

Marjan Fadavi (Economist)
Rachel Gicquelais (Epidemiologist)
Jose Coronado (Economist)
Freya Casier (Economist)
Adrian Soto (Physicist) --
Martina Balestra (Management...Scientist?)
Madison Hart (Social Sciences) (madisonkhart(at)
Carlos Viniegra (Economics)
Ella Jamsin (Physics, Sustainability)
Hugo Serrano (Computer Science)
Valérie Reijers (Ecology)
Damon Frezza (Operations Research)
Markus Junginger
Carla Rivera (Ecology)
Mark Kirstein
Horacio Marchand (Management..Psychology..Mythology) You?

Bipartite Ecological Networks


Characterizing and constructing bipartite ecological networks (using machine learning or something else?) What are the differences in structure between networks of different interactions? How do nonbiological factors inform network shape (weather, location etc.)? How do they stack into a multiplex network? Can we devise a way to create realistic networks? Given a bipartite network can we determine which type of interaction they are composed of? Can we categorize bipartite networks from other fields (like economics, physics?) as most like one of these ecological networks and then use insights for both field?

Networks, metadata, and references available at Mutualistic

  • 34 plant-seed disperser
  • 143 plant-pollinator


  • 4 plant-herbivore
  • 51 host-parasite

You can email me at (kaylasale(at) to chat about ideas, feasibility, whatever!

Interested Participants

  • Alje van Dam (thought a bit about country-product, occupation-city networks and their applications)
  • Surendra Hazarie :
  • Deepa Rao []
  • Tom Logan
  • Anand Nair
  • Ximo Pechuan

Self-organization of cities


spatial patterns, self-organization, power-laws, traveling salesman


This project aims to look for patterns in the spatial arrangement of human settlements. Previously it has been shown that the size of US cities follows a power-law distribution, but is this also the case for other countries? If we for example look at the size and spatial arrangement of cities in the Netherlands the pattern seems to be quite different. In this project it would be cool to look at the spatial organization of cities.

Possible research question: - Are there differences between countries in terms of their spatial arrangement (more clustered vs more dispersed?) How does this relate to the traveling distance between cities? Does the traveling distance follow for example a power-law? How does this interact/correlates with the size of cities. - Is the spatial arrangement dependent on the environment? For example if a countries has less favorable places (heterogeneous environment) do we also observe more clustering (adaptive strategy)? How does the interaction between facilitation (more resources if together) versus competition (disease outbreaks etc) influence the size and spatial arrangement of cities (and how is this influenced by the environment)? - Can we observe differences between the time the areas were colonized? (Australia and US as immigrant countries vs european countries?). Are cities in 'older' countries more likely to be dispersed (means of transport at time of settlement influences the distance between cities). Older cities more likely to have 'memory/legacy' effect of competition?

If you are interested let's get together and meet me or Tom or email: valeriereijers(at)gmail(dot)com.

Interested Participants

  • Tom Logan
  • Alje van Dam
  • Aaron Schwartz

Delayed Coordinate Mapping and Predicting Epidemics (or other time series)


The goal of this project will be to learn how to use coordinate mapping and other dynamic nonlinear models for predicting time series data. Many people have had some interesting ideas like (please feel free to add your own it's not here):

- How do predictions from a mechanistic model compare with those from a dynamic model and which is better?
- Similar to above, generate a null model of noise that resembles the data, and see if coordinate mapping claims to infer something from this vs. the real data (Josh tells me there's some literature about this he could point us to).
- Evaluating the type of disease dynamics these methods could be used on (e.g. seasonal epidemics, persistent infections, sporadic ones, etc).
- Can you use these methods to predict disease epidemics at all, and what are their limits?
- Conceptual/theoretical question: does the relative performance of nonlinear time series analysis and conventional epidemic models depend on collinearity between predictors and the role of unmeasured confounders


This paper is a good place to start our discussion:

Lunch Meeting

Let's meet today (June 14th) during lunch to discuss project ideas!

Interested Participants

  • Spencer Fox (
  • Tim Kunisky (
  • Shing Zhan (
  • Rachel Gicquelais (rgic (at)
  • Hilje Doekes (h.m.doekes (at)
  • jake weissman Ijw4336(at)
  • Alicia Kraay (amullis (at)
  • Shanee Stopnitzky
  • Elaine Bochniewicz

“Can complexity science save the world?”: a one-hour journey through open-minded discussions


“Can complexity science save the world?” At dinner today, Carlos Viniegra proposed this question jokingly, mainly to mock my (Yao's) oversized project questions.

But hey, why not start with this rather silly question, take a one-hour journey of open-minded discussions, and see where we will end up? One can certainly ask this question to any number of scientists, and have a ton of fun discussions. But we all know that the whole is greater than the sum of its parts, so the discussion ought to be even more fun as a group activity!

Come and share your deep concerns of the world and humankind, the “Holy Grail” questions of your research, the annoying limitations of your methodology (the ones that got swept under a rug)? Maybe through the diverse real-world big questions, theoretical questions, and methodological questions, something interesting will emerge, and innovative solutions can follow. And here is the bottom line: what other chances are there to hear about ideas/biases/rants from such a diverse group :) ?


Time: Tentatively this Friday, June 16, after the SFI BBQ, and over beer
Discussion moderator: Carlos will moderate! He will apply moderation procedures to stimulate innovative ideas.
Bring: Your rants and laughters, and maybe your world-saving superpower (optional, although JP said we all have it)

Interested Participants

  • Yao Liu
  • Madison Hart
  • Alex Jurgens
  • Andres Munoz

The Formation and Evolution of Religious Communities


network analysis (stochastic block modeling?), agent based model of diffusion, social networks


This project approaches the emergence and diffusion of human religious communities through inferential historical methods. The early history of Sufism, an Islamic religious movement which began in the ninth century CE in Baghdad, is not well understood, although there exists abundant data to study the interactions of its seminal members. This project attempts a reconstruction of these formative processes, using a relational dataset consisting of significant figures from different religious tendencies from between ca. 800-1000 CE and the directed connections between them. This data is derived from medieval Arabic prosopographical literature related to three different religious tendencies in premodern Islamic Iraq: traditionism (the transmission of hadith, the remembered sayings and actions of the prophet Muhammad [d. 631]); ascetics; and Sufis.

Each of the approaches to this data attempted in this project uses inferential methods to reconstruct the following dynamic historical processes 1) the membership structure of Sufism as compared to other groups; and 2) the diffusion of membership based on descriptive categories. Details of these discrete questions are discussed below:

1: Network Topology As of yet, there has been no attempt to reconstruct the network that connected Sufis in Baghdad to one another, and with their traditionist and ascetic contemporaries. Is it possible to leverage existing data to describe membership in these three tendencies? In this section, we describe the topology of the Islamic Baghdad in terms of agents that were traditionists, ascetics, Sufis, or some combination of these. By dividing the dataset into time steps, it is hypothesized that the analysis of initial developments will enable the successful prediction of the network position of actors–that is: traditionists, ascetics, Sufis, or some combination of these categorieş–in the final time step.

2: Diffusion Model In many cases, modern historians uncritically accept the categories that later historians used to describe collectivities. Is it possible to describe the evolution of groups "from the bottom up"? In this section, we use a diffusion model to describe the growth strategies (e.g. R, K) of the religious tendencies, described above, to measure the the capacity of a given actor to serve as an effective transmitter of one or more of these tendencies.

If you are interested please HMU: jeremy.farrell (at) emory (dot) edu.

Please add your name below if you'd like to get a piece of the action. 1 Uzay Cetin (uzay00(at)
2 Makoto Jones
3 Valérie Reijers
4 Nayely Velez-Cruz
6 Başak Taraktaş
7 Jeremy
8 Hilje Doekes

Possible Participants
7 Marjan
8 Alberto
9 Martina

From Makoto- I have an idea that relates to this, I think, that comes from the realm of antibiotic stewardship. I would be interested in doing some analysis of religious networks because I think at some point I could apply to my field. The antibiotic stewardship idea is that bacterial genes and human memes travel along related but not identical networks. In order for human memes to not die out, presumably they need to be reinforced to build critical mass. If some practice becomes monolithic and rigid then it can lead to collapse as soon as a gene for bacterial resistance arises that can move across the network unimpeded. Dominant antibiotic-resistant bacterial strains are similarly vulnerable to an attack. So in a religious context, I would be interested in what predicts branching in terms of adversarial competition. Maybe like a network implementation of Conway's game of life.

“Memory” in biological, ecological, social, and cultural systems


“Memory”, or legacy effects, are prevalent and significant in many worlds. Some examples in different systems include:

  • Vegetation that previous experienced drought might respond to water limitation differently.
  • Archeological sites (indicators of human activities in the past) are not always found in the most logical or optimal locations. Same goes for modern city developments.
  • Social practices (e.g. agricultural practices) are often driven by legacy, not optimal productivity or quality.

Potential research questions

  • Are there general ways to quantify characteristics of memory (e.g. the length and strength) across a diverse set of systems (biological, ecological, social, cultural, etc.)? Perhaps statistical, machine-learning approaches?
  • When we take memory into consideration, do we improve our prediction of the system’s behavior?
  • Sometimes, the “initial condition” of a system is unknown. Can we infer a system’s history from its subsequent behavior?
  • Under different regimes of perturbation, do memory help or hinder the function (survival, resistance, resilience, stability, etc.) of an organism, a community, a practice, a culture, etc.? The answer is likely different for a system versus for individual entities within that system.
  • What is the role of memory in its diverse forms (be it adaptation to the environment, learning from past experiences, or continuing/updated utilization of past infrastructures)? Can we view memory as a fundamental trait, on which natural selection occurs via evolutionary processes (in a broad sense)?

Interested Participants

  • Yao Liu (Please send an email to when you join the group, or join the #memory channel on Slack)
  • Jake Weissman (jw4336(at)
  • Alex Jurgens
  • Marco Pangallo
  • Yael Gurevich
  • Shanee Stopnitzky
  • Mark
  • Sean Wu

Cross-cultural Wikipedia Analysis


Wikipedia, cross-cultural, network analysis


Wikipedia is made up of user-generated content from people around the world. There are versions in 285 languages. Assume that users generate content in their primary language and that a person’s primary language has geographic or cultural correlations. Under that assumption, one might be able to tease out differences in cultural perceptions of an event/topic/etc. based on what is said, excluded, or linked to on the page. If you find this interesting, include your name below.

Interested Participants

  • Kyle Lemoi
  • Romona Roler
  • Hilje Doekes
  • Alex Jurgens
  • Elaine Bochniewicz
  • Madison Hart
  • Jiří Moravec
  • Rémi Lamarque
  • Katarina Mayer

Movie Data Analysis


Movies, modeling, complexity measures, generation and prediction


I was thinking about how people who study film often talk about "film language". Natural language is well studied (f not well understood) complex system, so I found myself wondering if we could apply some of the same methods to film. I found this database called Cinemetrics that has a large set of shot length per shot data, as well as all sorts of other data like shot framing, camera tilt, ect. I personally am interested in trying to model shot length data as a process and looking at entropy rate and statistical complexity as signatures of the data, potentially with the goal of being able to classify movies (action versus documentary, or by director, ect). There's also some interesting previous statistical work on the site that poses questions about psychology of editing, trend lines over time, and meaningful ways to find similar films. I don't think this has to be a huge project, but it seems like it could be a fun way to compare and contrast analysis skill sets across disciplines.

If you're interested please let me know! (Especially if you have experience in writing web scrapers.)
amjurgens AT
ajurgens on slack

Interested Participants

Alex Jurgens
Yao Liu
Jose Coronado

Autocatalytic networks

Growth, innovation, booms/busts, creative destruction

Still very vague idea; for people interested:
Pnas paper:
Interpretation as model of innovation:

Player Behaviour and Team Dynamics in a Competitive Online Environment


MOBAs, or multiplayer online battle arenas, are online video games that bring together random players to duke it out. These games generate tons of event data that can be utilized to test hypotheses about individual and organizational behaviour. There are outcome data that indicate victory or defeat and the performance characteristics of each player at the end of a game; there are also time series data capturing the behaviour and performance (e.g., gold accumulated and objectives accomplished) of each player throughout a game. For this project, I am proposing to analyze data from League of Legends (LoL) by Riot Games ( LoL is a real-time strategy game whose classic mode involves two teams of five players. Riot Games boasts 100 million monthly players, with millions of games played daily around the world.

Riot Games has made some of the data available via its API ( Using this awesome data, we can ask questions about player behaviour and team dynamics. What kinds of players are more successful? Passive or aggressive in terms of play style (e.g., calculated or yolo)? Jack of all trades or master of one (there are many different types of characters, e.g., marksmen, melee fighters, and healers)? Leaders who call good shots or those who comfort team members that are getting rekt? What makes a good team? A leader followed by supporters, or no single leader? What happens if there are two alpha-types in a team? If there are downers or douchebags who refuse to cooperate, toxic as heck, and/or intentionally do so bad that it is practically 4 versus 6? There is some recent literature involving the analysis of LoL data. We will review some papers and then brainstorm.

The bigger question is whether we can apply whatever insights we learn to better understand other competitive and/or virtual environments (e.g., sports, hackathons, and virtual reality social platform), to build stronger teams in organizations, and to become better team players ourselves.

Things we might be doing:
1) Build a machine learning method to classify players (e.g., selfish versus team-oriented) based on time series game performance data (e.g., number of champion kills, number of shots called, and gold earned)
2) Using the labelled data generated from step (1), test various hypotheses about (a) individual behaviour (e.g., are players who are more willing to take one for the team more successful in the long run?) and (b) team dynamics (e.g., are teams with mostly adaptive players more likely to win?)
3) Explore the dynamics underlying the formation of a competitive ladder (in competitive mode, there are various ranks, i.e.,, bronze, silver, gold, and so forth), which may shed some insights on how some people get more ahead than others in real life
4) Chaotic dynamics of player and team behaviour (there are time-series data of character positions on game map over time)

This project involves:
1) Non-linear time series analysis (e.g., how do we model time series data for 10 players which interact with each other to measure how one team works better together than the other?)
2) Unsupervised classification (e.g., how do we cluster the time series data to assign attributes to players?)
3) Multi-layered networks (e.g., network of different types of players dominating over each other)
4) Evolution of networks (e.g., LoL has frequent patch updates that often significantly alters game play and causes imbalances between play styles. How does the player-dominance network change over patches? Maybe if one imagines the characters as biological species and the patches as changing environments, then one can model it as an ecological system.)
5) Game-theoretic models (e.g., are less selfish players more successful in climbing the ladder?)
6) Agent-based models (e.g., can we predict where a player ends up over given some set of attributes?)
7) Chaotic dynamics (e.g., what are chaotic variations of player strategic movement over time? There is one ultimate attractor, which is destroying enemy base, and a few transient ones, which are sub-objectives that periodically appear throughout the game)
8) Information spread (e.g., Players can ping each other to notify team members of incoming threats or upcoming objectives. How can we model how players react to pings?)

I think there is something for everyone. If interested, please give me a shout at or join the Slack channel #game-on.

Interested Participants

  • Shing Zhan (
  • Abdel Abdelgabar (yomegaland AT gmail)
  • Bleu Knight (happylittleaccident AT gmail)

Communication and Cryptography with Chaotic Maps


chaotic maps, information theory, information transmission, cryptography, secure communication


Chaotic maps were discussed today (June 14th) during Liz's talk on dynamic systems. There is a whole zoo of known chaotic maps (there's a pretty good list on wikipedia) and in general if you can find a generating partition for a chaotic map you can use symbolic dynamics to capture the dynamics of the continuous system with a discrete sequence. This allows for easy computation of information theoretic quantities like entropy rate, redundancy, excess entropy, ect. I would be interested in thinking about this sequence as a coding channel for a few different maps. It's also possible to use chaotic maps as a cryptographic scheme (encoding images, for example) and it would be fun to take a look at new approaches and feasibility in these methods, maybe do some information theoretic analysis. Some of this might already have been done, so this group could also be a bit of literature review on the current state of chaotic dynamics.

Sort of analysis I'm talking about:

List of chaotic maps:

Other related papers:

"Introduction to Symbolic Mechanics
"Symbolic Dynamics on One Dimensional Maps"
"Symbolic Dynamics of Noisy Chaos"
"Secure communication using Duffing oscillators"
"A Canonical Partition of the Periodic Orbits of Chaotic Maps"

If you're interested please let me know!
amjurgens AT
ajurgens on slack

Interested Participants

Alex Jurgens
Jingnuo Dong
Chris Miles
Yao Liu
Jose Coronado
Shing Zhan

DNA of Capitalism

More serious title: Are multiplicative dynamics of wealth of the rich part of a countries' population a signature of capitalism?


One of the first power laws ever documented is the power law behaviour for the upper part of the wealth distribution of countries (Pareto 1897). Since then power laws in wealth distributions seem to be a well documented fact ( and are regarded as a universal feature of economies (Yakovenko and Rosser 2009). The lower part of the wealth distribution is well approximated by a Boltzmann-Gibbs distribution and generated by additive dynamics of the process, which is what happens if wealth is mainly accumulated via labour income. However, the generating mechanism for the upper power law part is a multiplicative dynamics, as the wealth of the "richer" part is predominantly accumulated not via labour but compound interest from investments like equity, fixed incomes, real estate etc.
Now several questions pop up. Is multiplicative dynamics of the wealth of the richer part of the population only a signature of capitalist economies? So far, I am not aware of any data which presents analyses from more or less socialist economies/countries like e.g. former GDR, former Soviet Union, China, Venezuela, Cuba, Vietnam, North Korea etc.

Skills needed

I guess without a good level of the languages of the mentioned countries it's very hard to dig out data from the institutions. Hence, participants with good knowledge of German, Russian, Chinese, Spanish or Korean will be valuable. Additionally, knowledge on power law generating mechanism is valuable, which includes knowledge about stochastic processes. Furthermore, we will have to handle probably big amounts of data and need to be able to properly estimate and distinguish between exponential, logarithmic or power law decays in the tail of the wealth distribution.

If you find this interesting, include your name below.
If you have any questions or ideas just approach me (mark.kirstein[at] as some already did or come to the first casual meeting today on Wednesday June 14th at 16:30 in the main lecture hall. Further contributions and ideas to expand this project are warmly welcome!

Interested Participants

State Capacity with agent-based modelling / Conflicts and Evolution

Key words: tax, world-systems, migration, agent-based modelling, class.


State capacity is a key word in the international society. Generally speaking, it involves the ability of one state to achieve certain goals. One crucial part of state capacity is to collect enough tax so that the state could work effectively. Then the questions will be: are the people willing to pay tax? And are the state officials willing to work hard? If we introduce the multi-state system, it could be even more complex. If the people in one state are taxed too heavily, they would migrate to other states. And some strong states can require tributes from some weak states with war threats so that they can buy off their people.

We can work together to model this dynamics!


1. Growing Artificial Societies- Social Science from the Bottom Up ( by Epstein, Axtell)

Interested Participants

  • Junfu Zhao (
  • Sheri Xue Guo (
  • Zhiya Zuo

Organic Boundaries/Cluster within Cities


Our cities are often divided into many different districts or areas; these areas are usually for urban planning, urban management and administration purposes. These boundaries are fundamentally artificial.

In reality, the city has its own way of organizing its function and activities, e.g. a facility may attract people from not only this area, but from the surrounding areas; a few estates may be very well integrated due to certain reasons.

So in that case, the city would have its own organic, self-organized boundaries/cluster. If we can develop a way to identify these boundaries, they would be very helpful to the urban planning and management, e.g. gain better understanding in the city's spatial patterns, determining the optimal distribution of resources and facilities. The project intends to analyse Singapore as a case study with real dataset.

Interested Participants

  • Zhou Yimin (
  • Xue Guo (
  • Sean Wu (
  • Doheum (
  • Alicia(
  • Xinya (
  • Tom Logan
  • Kyle Reing (
  • Yuji Saikai (
  • Huang Tang

Network's evolution: response and spread


Through discussion with people, we've seen similarities in the spread of information (e.g, religions, languages, diseases) on networks of people. We would like to explore in more abstract sense if the spread of information follows similar patterns in every kind of complex system. The aim of the project would be to create an abstract model of information spread that could apply to a variety of fields.

Interested Participants

  • Spencer Fox (
  • Rémi Lamarque (
  • Yael Gurevich (yael01[at]
  • Elliot Nelson (
  • Huang Tang
  • Rachel Gicquelais (rgic[at]
  • Alicia Kraay (amullis [at]
  • Nayely Velez-Cruz (

Agent Based Modeling and Delay Coordinate Embedding


Often cases in my work, you have very little or no data. To get an idea for the behavior of they system, we often create agent-based models to get a feel for what the real system would do. Without data, we cannot calibrate the parameters of the ABM or validate the results. This leaves us to doing sensitivity analyses and uncertainty quantification in order to give our customers confidence in the results. I think Delay Coordinate Embedding could be a useful tool in performing additional analyses on the results of our ABMs. The main idea is to use the output of an ABM and treat it as a non-linear time series data. Then see if we can use DCE to get information about the dynamics of the "data". My goal is to show proof of concept for an ABM that outputs data where DCE concludes chaotic behavior, but just from analyzing the results of the ABM you might now have seen the chaotic behavior. I am not saying this will always give additional information to the results of an ABM, but I think it is a tool we can add to our toolkit to evaluate the performance of ABMs.

Meeting details 6/16/17 - We decided to go forward with two examples: a 4 species predator prey model and an economic model. Next steps are implement ABMs for each model and to identify parameter regimes where we think the systems will exhibit various behaviors (ex - stable equilibrium, periodic orbit, and chaos). Meghan and Stephan decided we will try to implement the ABMs in NetLogo since we have never done it so it will be a good skill to learn.

Next meeting Monday 6/19/17 at lunch

Questions for the group: What is the best way for everyone to file share? This will allow us to add to the final report as we go to save time in the last week.


Interested Participants

  • Meghan Galiardi (meghangaliardi (at)
  • Tom Logan
  • Shanee Stopnitzky
  • Marco Pangallo
  • Yuji Saikai (
  • Markus Junginger (
  • Stefan Bucher (sfb311(at)nyu(dot)edu)

Persistent Homology of Fitness Landscapes

Motivated by the lecture, apply topological data analysis to fitness landscapes in a network representation. Contact: Ximo

Defining simplicial complexes for directed networks:

Statistical Mechanics of Ecosystem Assembly

Using stochastic processes improve on the approach


Interested participants

  • Valentina Baccetti

Effects of network topology in business information processing and dissemination mechanisms or: How to avoid financial and market death by PowerPoint and foster information worker productivity.


As automatization has accelerated within the manufacturing environment, sophisticated companies now know, to the last bolt and second of labor spent, the full cost of their manufactured goods, thus, fostering cost saving strategies and increased productivity. At the management layer, the wide dissemination of digital technologies such as email, PowerPoint, corporate social tools, and management methodologies (pmbok, etc.) presumably have also improved productivity. However, given the complexity and subjectivity of information work, little effort has been done at understanding the emergent phenomena (costs, productivity and quality changes) created by the amalgamation of traditional management structures, such as hierarchical (tree) management structures, and the before mentioned technologies. It is also probable that computerization, by changing variables such as the noise/signal ratio, has created the opposite effect over information worker outputs (more cost, less productivity and quality)…

Kickoff meeting June 15th 18:30 at the cafeteria, blackboard area

Follow link for draft of scope document:!AtgTQ5zzPp5umwkLc2cK3wbBQpXQ

Interested Participants

  • Stephen Leese
  • Markus Junginger
  • Damon Frezza
  • Horacio Marchand
  • Carlos Viniegra

Training a self-organizing agent-based model (motivated by Albert Hubler's ball bearing experiment)


self-organization, maximum entropy production.


The goal is to recreate something similar to Albert Hubler's ball bearing experiment with Tetris training --- but within an agent-based model. The first task is to create an agent-based model that is capable of self-organization in response to increasing 'throughput'. By using this throughput, we can reinforce the ABM when it is performing well at a task or game. An ABM may allow us to better understand the principles of self-organization and maximum entropy production by the ease of experimentation in simulation. In addition, the hope is to give insight on how to use these self-organizing principles for the means of training an ABM.

Interested Participants

  • Chris Miles
  • Kyle Reing
  • Katarina Mayer

Data Sets

Some datasets that might trigger an idea


bunch of USA data (towns, cities, states, education, skills, occupations, industries...)


historical database of 'globally known people' with place and date of birth and death (not sure if it's publicly available though..)

Ethnographic Atlas

  • database of 1167 societies and their ethnographic descriptors (see here )
  • Corrected ethnographic atlas: here
  • some links might be dead, so here is reupload of R version: here

Publicly Available Time Series Data

Archived Projects "The Parking Lot"

Human Movement Dataset


Dataset looking for a research question. Data is a time series of Linear and angular X, Y, Z of the upper limb of 25 different people completing a series of activities.

  • One possible thought, is there a way to recognize the difference between the way a control (normal) moves versus a stroke survivor?
  • Is there a way to identify the severity of an injury based upon movement patterns?
  • Can specific movement patterns be identified consistently?

Interested Participants

  • Elaine Bochniewicz