A s one of the early cohorts of The Data Incubator, I was introduced to a number of machine learning techniques in NLP and Sentiment Analysis. Taking my background into consideration, I built a simple web app that looked at the relationship between stock sentiment and US Equity performance.

Here's my video demonstration of this guy:

This webapp was hosted on a Digital Ocean Ubuntu server and built on Python Flask with Plot.ly graphics. The machine learning used pandas for data manipulation and scikit and ntlk for corpus extract and analysis. The data itself came from JSON files provided by StockTwits, a platform for investors and hobbyists to talk about trading and portfolio management.

Note: I took the webapp down shortly after the program ended, since much of the work was fairly primitive and a lot of other analysis could be done with these tweets, but hope to resurrect it one day!