For my thesis I learnt to code in Python. John Stachurski and Thomas Sargent’s quant-econ site provides a great introduction to Python for economists. Python is a joy to use, but it’s a bit slow. Here I rely on Cython which is basically a Python to C translator, see this post.

Recently I’ve been working on adapting some machine learning methods to economic problems…


econlearn is a python machine learning toolkit, based on the methods I developed for my thesis. econlearn provides some reinforcement learning (i.e., approximate dynamic programming) and supervised learning (i.e., non-parametric regression) algorithms suited to economic problems. That is noisy problems, requiring large sample sizes, but involving relatively few dimensions. In particular, it implements a batch version of Q-learning using tilecoding for function approximation.

You can download econlearn from github here. The code and documentation remain a work in progress. If you have any problems with it please let me know.