Live Government Policy Simulator

When a government invests into a social program, the program might not generate benefits for years. This is not only because the program may require time to cycle up into operation, but also because its very goal may be to put people onto paths to success later on. Sending kids to preschool, for example, has been shown to give them an earnings boost decades later as adults.

The above simulation visualization depicts high school students stepping their way through a very basic version of the college application process. And in the make-believe dataset to which this visualization is attached, a fair share of the students unfortunately do not make it.

As the operator of the simulation, you get to decide just how much to invest into social programs to try to help--programs such as after school tutoring. The simulation is attached via JavaScript to a nonlinear predictive time-series model, which determines how the students respond over the course of their schooling to your interventions. These responses appear visually.

One note: in principle this model could have been fitted using real government data, though it wasn't (I made up some data for it).

Also, you may notice that the students don't respond immediately to your interventions--one of the reasons why making good policy can be difficult!

Try these things with this demo:

  • Click "take a tour" -- the tour lasts around 90 seconds
  • After this, start your own simulation
  • Experiment showing and not showing people who started high school either on or off track
  • Implement tutoring interventions -- and watch the people who receive them, now surrounded in yellow
  • Implement extracurricular interventions -- and watch the people who receive them, now surrounded in a darker color
  • See an alternative visualization that represents the same data and model, but in a different format