— In the News —
When it’s possible to record the exact movements of players in team games such as football and basketball, how can algorithms crunch this data to provide meaningful insight?
Insightful article about common problems faced by teams of data scientists and engineers and how Stitch Fix managed to build A Better Way to structure a data science department.
This report by MIT Sloan Management Review and SAS finds that most companies are not prepared for the changes that are required to achieve success with analytics. Along with exploring why, the report also examines the characteristics that lead to success.
AlphaGo vs Lee Se-dol
The recent Go match between Google's AlphaGo and Lee Se-dol, one of the world’s top players, is old news already but what an exciting match! Everyone seemed to be covering it but here are the must-reads:
— Tools and Techniques —
This is a reference implementation of the software described in DeepMind's 2016 Nature publication, "Mastering the game of Go with deep neural networks and tree search." This is not yet a full implementation of AlphaGo but it looks like a fantastic project to get involved with.
Leaf is a Machine Intelligence Framework that's been engineered to bring modularity, performance and portability to deep learning. Ultimately, the goal is to bring real-time, offline AI to smartphones and embedded devices.
datacleaner is a Python tool that automatically cleans datasets and readies them for analysis.
Nice way to approach a JSON data file that's too big to fit into memory.
— Resources —
Using practical examples, this book shows how to create reproducible workflows, including data gathering, analysis, and presenting results. It's free to download or if you prefer print, it's also available from Amazon.
— Data Viz —
Make your own shot charts for any NBA player dating back to 1996! In addition to showing shot locations on the court, this open-source library calculates aggregate statistics and compares the selected player to league averages at different areas of the court.