SayWhatYouMean is a project that aims to provide users with an interactive, visual representation of an individual Twitter handle.
It does this by leveraging NLP (Natural Language Processing), Sentiment Analysis, and CUSUM change detection.
The most frequently discussed topics (hashtags) are first summarized to show the distribution of tweeted sentiment. An individual topic can be filtered further to show the individual tweets over time vs. the general population sentiment.
Finally, CUSUM change detection is used to identify key areas of change in population sentiment, choosing a threshold where a change would be identified through the use of a slider.
