web application for text mining / sentiment analysis
Small webapplication that provides insight into industry and company sentiment.
Data has been gathered from twitter and various online news sources.
Sentiment scores (0-1) have been calculated based on a stochasticd gradient classifier (logistic modell) which in turn used a TF-IDF Vectorizer as feature extractor.