Andrew Reece: Data Science
These are repos for different data science projects I've worked on in professional and academic capacities.
In many cases, links to websites and screencasts are available from the landing pages of each repo.
Also see my website for a summary of my work, along with links to my CV and blog.
Lead Poisoning Prevention
Developed a predictive model and interactive data interface for Chicago Department of Public Health, to aid in city-wide prevention of lead poisoning
* only a D3 code snippet is available, due to privacy constraints
- Innovation Mining Used NLP, LASSO regularization, and permutation modeling to algorithmically detect differences in innovative thinking
- SEE-Net Developed a network analysis web app (D3, Python) to model the evolution of social networks
- Barbaric Yelp - Created simple collaborative filtering recommendation system a la Yelp
- Bayesian Tomatoes - Used Naive Bayes classifier/text vectorizer to predict pos/neg reviews on Rotten Tomatoes
- Governor Races - Made bootstrapped predictions of future gubernatorial elections based on previous polling & error data
- Monty Hall - Algorithm reproduces Monty Hall problem with N doors, N trials
- Machine learning/NLP for sentiment & popularity prediction in New York Times articles
NYT project website
- Predicting Obama - Different weighting schema for computing electoral college votes via polling data for 2012 US Presidential elections
- Senate Voting Network - Modeled partisanship in US senate voting records using NetworkX & Gephi