Social Simulation 2018 in Stockholm

Last week we attended the Social Simulation conference in Stockholm. It was the 14th annual event organised by the European Social Simulation Association that work to promote social simulation and computational social science. The theme of this year’s conference was: looking in the mirror; reflecting on a growing field and community. There were about 120 attendees who had come in from the far reaches on the world: from Colombia to China and Australia.

Our purpose for being there was to engage the community and to discuss our newly released (closed beta) cloud-based platform, Sandman. Sandman is a cloud-based platform for doing scalable, agile agent-based modelling. For more information, see the website that includes our whitepaper and a number of case studies.

We arrived Monday to get setup: we’re given a “booth” (a corner by the coffee area) to be based – from which to demo our technology. During coffee breaks we ensnared conference goers to show them Sandman: “look how easy it is to run your models across 1000s of cores in the cloud!”

The conference kicked-off for real on Tuesday – but we joined workshops on Monday afternoon.

• Tackling prediction in empirical agent-based models.
• Using theories on human-decision making for in ABMs.

There were three very interesting keynotes:

• Bruce Edmonds who talked about context, how it is unavoidable, but mostly ignored in models.
• Milena Tsvetkova discussed various research projects, in particular, empirically validating the assumptions underlying Schelling’s famous model of segregation, and looking at the emergence of inequality in social groups.
• Julie Zahle presented philosophical arguments about methodological individualism and how ABM are not necessarily tied to individual explanations but also holistic ones.

From Tuesday to Friday afternoon there were sessions. Here are a selection of our favourite tracks from the week (for more info, see the conference programme):

• Education advancements in social simulation
• Modelling social science aspects of socio-ecological systems
• Social simulation in the policy world
• Validation of agent-based models
• Computational economics: using simulation-based techniques to reflect on economic theory
• Model sharing, code sharing, and interdisciplinary collaboration in large-scale simulation models

The sessions reflected the breadth of work being done across a number of areas: from sociology to economics and even archaeology. As for tools used to develop models, NetLogo is clearly still dominant with some Java and Python, and even a few using R and Julia.

The social simulation community is very diverse, bringing together people from a range of backgrounds and interests. This certainly made for a fascinating week, seeing the wide array of work being done and discussing our technology with participants.

Thanks to everyone for being so welcoming. Hopefully see you next year in Mainz, where the theme will be policy modelling.