Scaling Open Source Policy Models and the Biden Plan

Hugo Bowne-Anderson August 30, 2020

, , ,

Richard Evans, Advisory Board Visiting Fellow at the Baker Institute for Public Policy at Rice University, joins Matt Rocklin and Hugo Bowne-Anderson to discuss open source policy modeling in Python, the power of Dask, and the Biden Plan.

Opening with a discussion of why models of public policy should be open, we’ll then jump into the Policy Simulation Library, a collection of open source models and data preparation routines for policy analysis.

We’ll then take a deep dive into a model of the economy of the United States that allows for dynamic general equilibrium analysis of federal tax policy and leverages Dask for distributed compute (in four different places in the code base!). We’ll see how this model can be run and scale locally and also built into a web application. We’ll also check out the current results analyzing the Biden plan.

After attending, you’ll know:

  • How the open source Python stack can be leveraged to built simulations of economic policies
  • How Dask can be used to parallelize such simulations, both on a user’s local machine and on a web application server (Google Cloud, in this case)
  • How to reason about the importance of public policy models being open.

Join us this Thursday, September 3rd at 5pm US Eastern time by signing up here to dive into the wonderful world of scalable public policy simulation in Python!

A graphic for Coiled's Science Thursday with Richard Evans ("Scaling Open Source Policy Models and Analyzing the Biden Plan").

Ready to get started?

Create your first cluster in minutes.