My always expanding collection of data science related projects plus smaller tidbits for fun and learning.
Ongoing work to forecast weekly sales for a local, boutique retailer. This is part one of a larger project to forecast product inventory and automate re-stocking orders.
A fun Sunday pre-game exercise to scrape the current player rosters for each team from nfl.com and classify Green Bay players as linemen or wide receivers based on height and weight using logistic regression. A great MIT Opencourseware video inspired me to do this. I was really surprised that there are almost 2,000 active players in the NFL.
Gradient descent is how you minimize cost functions in everything from linear regression to neural networks used in deep learning. Since it's so important, I thought it would be worth the time to learn and implement it from scratch again.
A large and growing collection (mostly Jupyter Notebooks) with fun exercises, labs and mini-projects from sources like Harvard's CS109 online class, etc. Notebooks on everyting from data wrangling, SQL, EDA, hacker statistics, classification, regression and unsupervised learning are here.