Last year, I open sourced my first library, PyMC3 Models, which features custom PyMC3 models built on top of the scikit-learn API. This talk is about some of the things I learned along the way.

The first half will focus on lessons learned while writing my library\: - The importance of just starting: releasing even a small bit of work is worthwhile because it could help someone. - Using toy examples: implementing a basic linear regression helped me improve the broader structure of my package. - Documentation: Spending time on documentation was tedious in the beginning, but paid off in the long run. - Copying: I borrowed heavily from other open source libraries and that’s totally ok!

The second half will focus on being the sole maintainer of the library including: - how to handle normal contributions - strategies for dealing with aggressive contributors - taking care of yourself when you’re the only one working on a library

I hope this talk will encourage you to start your own library and give you some tools as you maintain it.