A Python program, performing sentiment analysis of top trending tweets, and setting my Philips Hue to a colour matching the worlds current emotion
hence a literal mood light đź’ˇ
It uses the NRC Word-Emotion Association Lexicon (NRC Emotion Lexicon). I have 6168 words, where some words have multiple emotions associated with them.
I use Tweepy, a Python Library for using the Twitter API. I get the top trends in the UK, and then for each trend, get 100 (which is the max Tweepy will return) tweets with that trend and analyse each one.
The top emotion corrolates to which colour the Philips Hue changes to:
| Emotion | Colour |
|---|---|
| Anger | Red |
| Fear | Red |
| Anticipation | Orange |
| Surprise | Purple |
| Sadness | Dark Blue |
| Joy | Yellow |
| Disgust | Green |
| Negative | Blue |
| Postitive | White |
- Add emoji analysis
- Find/Generate lexicon with more current/slang/topical words in (e.g. Brexit, fetch)
- Deal with negation (e.g. Not Happy, where as currently get Happy)