NLP Election

Analyzing each candidate tweets and replies they received, to predict 2020 presidential elections

• Please note beacuse of time limitation on this project, classification algorithum isn't included yet in this project which was initialized on an automatic approach but all sentiment analysis are covered

This project was build using:

• Liabraries:

  1. TextBlob to analyze sentiment
  2. WorldCloud to visualize data
  3. Pandas to read tweets in dataframe
  4. Matplotlib to visualize data

• Tweepy Installation:

  pip install tweepy
  

• Sign up for Twitter Developer account to generate necessary tokens & keys

---

Fetch tweets

• each candidate: using API feature "user_timeline" This feature has a limitation on the number of tweets requested up to 200 maximum.

• replies to each candidate: using tweepy feature "Cursor" This feature can iterate through 10,000 tweets per request maximum which will increase our analysis accuracy.

Create functions to evaluate Rational & Emotional analysis of each tweet

• Polarity & Context : to measure Emotional reactions, where 1 is positive, 0 is neutral, & -1 is negative.

• Subjectivity & Tone: to measure Rational reactions, where 0 is objective& 1 is subjective

• Emojis

• Irony and Sarcasm

Donald Trump tweets were:

• 39.5 % Positive

• 12.50 % Negative

• 48 % Neutral

• 140 times Exclamation Mark was used

Joe Biden tweets were:

• 50 % Positive

• 18 % Negative

• 32 % Neutral

• Exclamation Mark was used 4 times only.

  

  

Voters' favor Joe Biden over Donald Trump, their replies sentiment analysis shows:

• Positive replies scored 58.76 % for Joe Biden & 47.22 % for Donald Trump

• Negative replies scored 41.22 % for Joe Biden & 52.78 % for Donald Trump

• Users' replies to Donald Trump tweets

• Users' replies to Donald Trump tweets

---

References

• API USER Timeline Limitation
• Regex Cheatsheet
• TextBlob & sentiment analysis
• Calculating Polarity and Subjectivity
• Hvae Fun with TectBlob
• Twitter Docs-Cursor