October 23, 2018
Two science-themed talks, sponsored by Wayfair: https://www.wayfaircareers.com
Carlos Faerman, Python Meets Chemistry
In this talk, I will explain the use of Molecular Fingerprints to describe the properties of small molecules. I’ll share a Python script to compare molecules; search for similar molecules in large databases; and describe how this approach applies to drug discovery.
Graham Voysey, Writing Python for Reproducable Research
Python is the lingua franca of scientific computing, machine learning, and data science. Regardless of language, research code and scientific computing have special demands placed on the developer that are probably not well understood by other developers at large. At our peril, we write code, generate results, change our code, and then struggle to re-associate which version of our code generated what result. This leads to great annoyance, wasted time, and potentially serious consequences like retractions or production failures. In this talk, I will briefly discuss some reasons why these problems arise. Then, I will offer some practical steps that I use to minimize these problems that form the basis of some best practices for development in these areas. These steps help me consistently generate code that is easy to reason about and produces output that provides permanent provenance, while also making your code easy to share and disseminate. I will provide real examples taken from production code from both academic and industry settings. Finally, I will demo One More Thing: my personal turnkey solution to resolve most of the issues I raise, and seek community input on it.
Meetup link: https://www.meetup.com/bostonpython/events/253168754/