Simulating Social Networks

Predicting how information spreads online

Our work on SimPPL is deployed with global newsrooms, with details available here.

SimPPL: Social Network Simulation

Here’s a Mission and Impact statement of what we’ve achieved!

Simulating a social network using agent-based models of information spread online.

I am building SimPPL - a social network simulator to demonstrate how to combine heterogeneous datasets in a principled manner so as to create an expressive model of online social networks that is conditioned on real-world data. It is part of my ongoing research on misinformation control to highlight the applications of such a tool towards understanding the diffusion of information and the evolution of beliefs on platforms like Facebook and Twitter. There are a rich set of downstream applications of such a simulator, including interventions to curb misinformation spread, and the causal modeling of online user behavior.

  • I received an independent Google Cloud Research Grant to support SimPPL.
  • I received a few community grants from Algovera AI to support SimPPL.
  • SimPPL has been accepted to be part of the NYU Tech Venture Workshop, 2022!
  • My team, SimPPL, is part of the NYC Media Lab’s AI and Local News Challenge!
  • I am collaborating with the Torr Vision Group at Oxford on the applications of such a simulation tool on estimating the effects of coordinated inauthentic behavior on content recommendations in social networks!