https://medium.com/@franckjay/bayesian-modeling-for-ford-gobike-ridership-with-pymc3-part-i-b905104af0df

Bike shares are a large part of the transport equation for cities around the world. In San Francisco, one of the major players in the bike share game is Ford with its GoBike program. Conveniently, they kindly release their data for people like me to study. I wonder if it is possible to easily forecast the ridership of the next day in order to ensure enough bikes are available to riders, based on past information?

This would be a fairly trivial task to complete if I were to use sklearn to build a Linear Regression model. Often I find myself looking for data sets to learn a new tool or skill in Machine Learning. I have been trying to find an excuse to try one of the probabilistic programming packages (like PyStan or PyMC3) for years now, and this bike share data seemed like a great fit.