This git repo contains all relevant data files, reports and graphics pertaining to my Master's thesis for Columbia University's Quantitative Methods in Social Science program. My project focuses on comparing the difference between sentiment analysis of text and emoji in a single tweet.
Hypotheses
- There will be a significant difference between sentiment analysis of text compared to the sentiment analysis of emojis.
- That in general tweets referencing Republican candidates (e.g. Trump, Cruz and Rubio) will contain more negative sentiment for emojis than tweets referencing Democratic candidates (e.g. Clinton and Sanders).
- That tweets that mention Democratic candidates are more likely to have higher negative text sentiment than tweets mentioning Republican candidates.
Want to see more?
- Check out Notebook.md for a daily log of process.