This repository is a playground to apply NLP techniques on twitter threads. Currently there exists a class that gets the last 200 tweets from any twitter handle and writes it to a database. This is then picked up by a python file which plots the twitter metrics to a frontend server using Dash and Plotly.

• Rename the cred_sample.py to creds.py and fill out the neccessary information.
  • You can set up the keys and tokens from: Twitter Apps (assuming you have a twitter account)
• start a pipenv environment this is especially useful for the Dash and Plotly server. You can install and use it from here
• One of the methods uses the ldamalletmodel which needs to be downloaded to a directory of your choice. This can be downloaded from here.
• Make sure you have sqlite3 installed on your computer.

• Apply NLP techniques using Spacy, Gensim and scikit learn about topic modelling [X]
• Generate user word clouds [X]
• Write twiiter data into a SQLite Database [X]
• Generate Dash frontend to display twitter metrics [X]
• Write a scheduler that inserts new twitter data into DB. [X]
• Create new model to analyse twitter sentiments, instead of using TextBlob.
• Get Like Factor for user.
  • Get collection of twitter replies and create a senitment polarity ratio (postive : negative replies) for the users tweets
• Get tweets by sentiment
• Av Time for Tweet response by followers
• Design most liked/ most retweeted and least liked tweet similar to twitter design.
• Twitter Len vs Twitter Engagement Analysis
• Deploy to Heroku
• Develop a Stock price for each user based on the like factor.