Outbreak simulator

reproduce simulator similar to this

Basic model

• Agents move

• Disease is transmitted between "touching" agents (distance less than some threshold)

• Agents are infected for 14 days

• At day 14, agents either recover or die

• Agents that move out of bounds are moved to the

• Out of scope: agents angle changes upon collision

Future features

• Ability to play with parameters,

  • plot stacked area of statuses over time with different features.
  • Compare through change in parameter at time x (e.g. number of people infected at day 100 or max people infected during a simulation, for different values of parameter (e.g. # static agents))

• Geo-plots

  • NYC taxi data set to simulate agents moving around city here
  • Mapbox to plot
  • Airline data
  • Restaurants here (What does shutdown do?)
  • Subway entrances (would be nice to get commute info)

• Use real data to model by city

  • E.g. transmission rates, population sizes. Can we model future rates?
  • Census data to get age distribution. Can we adjust recovery probability?

• Newly infected agents show as larger, or a different color.

• Calc probability of being infected as rolling average (e.g. 5 day average) of (# new cases)/(# uninfected)

• Run scenarios for centralized response/interventions: Test and Quarantine,

• Factor in hospital capacity: Set some threshold. If the number of active cases is higher than that, the mortality probability goes up.

Refactors

• Move underlying update functions to external script
• Host on server/ webpage
• split into server side feed built in Python (Django) and client-side app in Javascript (D3)