Skip to content
ML project to find the most similar account within top Twitter accounts
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
react-frontend
.gitignore
Procfile
README.md
Top10Accounts.csv
app.py
create_model.py
requirements.txt
runtime.txt
setupCSV.py
tweet_classifier.pkl
vectorizer.pkl

README.md

TwiSimilar

ML project to find popular Twitter accounts similar to a given account

Usage

Visit https://colinmarsch.me/TwiSimilar/ to try out the web app. After authenticating with Twitter the app will take your latest tweets and run them through the classification model developed, predicting one of the top 10 Twitter accounts for each tweet. The Twitter account that shows up the most frequently in the results is then returned as the closest match to your Twitter account.

Model

The code used to create the model used in the classification can be found in the create_model.py file. A LinearSVC model was used as the classifier.

Backend

The code for the Flask backend of the app can be found in the app.py file.

Frontend

The code for the React frontend of the app can be found in the react-fronted folder. This was initally created using the create-react-app npm package.

You can’t perform that action at this time.