TweetMood is a fascinating data visualization tool that analyzes that sentiment of tweets based on geographic location and influence.
In the demo above, the radius of each circle is influenced by both re-tweets for the unique post and follows for the general sender account. The color represents the sentiment, where the poles are pure green and pure red, indicating positivity and negativity, respectively. Opacity is an interesting aspect to the map. You can see overlapping circles of the same sentiment mark an area as more red or green.
We used machine learning research trained on IMDB data sets to detect the sentiment. Although the analysis is accurate, it's far from perfect.