Created by @sruthipsuresh Practice of Twitter Data Scraping using Twint and NLP Analysis. Inspired by this article, generated individual wordclouds for AOC, Pete Pete Buttigieg, Ted Cruz and Ben Shapiro in images folder. Also made subjectivity/polarity analysis and plotted in seaborn scatterplot.

See the wordcloud generated for AOC from her Tweets so far in 2021!

See the wordcloud generated for Pete Buttigieg from his Tweets so far in 2021!

See the polarity vs. subjectivity scatterplot of all the users listed above!

• Install dependencies in script.sh.
• Generate .csv files with tweets from usernames, TODO: configure userlist in modules
• Concatenate all .csv files in one (first strip header with sed 1d, then run cat command)

• Used Google Colab and referenced various articles to clean data.
• Subsetted to username, tweet text.
• Removed all Twitter characters/links from Tweet.
• Data cleaning for NLP analysis.
• Store cleaned data in clean.csv

• Followed tutorial on datacamp.
• Generated countplot.
• Generated test wordcloud using only one tweet.
• Combined all tweets in merged dataset for overall wordcloud.
• Combine all tweets by username and generate separate wordclouds.
• Save in images folder.

• Use dataset which is only processed to remove Twitter tags.
• Add polarity and subjectivity columns with Textblob, saved .csv
• Simple plot with username used as hue.