Skip to content

This is a simple python program which uses a machine learning model to detect toxicity in tweets, developed in Flask.

License

Notifications You must be signed in to change notification settings

mantreshkhurana/twitter-toxicity-detection-flask

Repository files navigation

Twitter Toxicity Detection Flask

This is a Flask web application that allows users to search for a Twitter user's recent tweets and get a toxicity score for each tweet using a pre-trained logistic regression model.

It has a Window GUI version which can also be used to do the same without opening the browser.

Table of Contents

Demo

Demo Video

demo.mp4

Screenshots

Light Dark
App Screenshot App Screenshot
App Screenshot App Screenshot

Pie Chart

You can see a pie chart which portrays the percentage of tweets that are toxic and non-toxic. It can be viewed by clicking on the view Pie Chart button which is located bellow following and followers count

API Keys Popup

After saving the API keys, you can close the popup and restart the flask server.

Light Dark
App Screenshot App Screenshot

Window GUI

App Screenshot

Installation

You may need to install some dependencies before running the program(some of the modules cannot be installed directly by using requirements.txt).

Get Twitter API keys from here.

Note: Make sure to apply for Twitter Elevated Access to use the Twitter API v2 endpoints, without elevated access you will not be able to use this app. (It's FREE until you reach the limit of 2,000,000 tweets per month)

To get started with this project, follow these steps:

Virtual Environment

I recommend using a virtual environment for this project.

git clone https://github.com/mantreshkhurana/twitter-toxicity-detection-flask.git
cd twitter-toxicity-detection-flask
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
touch .env
echo "CONSUMER_KEY=<your_twitter_api_consumer_key>" >> .env # replace <your_twitter_api_consumer_key> with your Twitter API consumer key
echo "CONSUMER_SECRET=<your_twitter_api_consumer_secret>" >> .env # replace <your_twitter_api_consumer_secret> with your Twitter API consumer secret
echo "ACCESS_TOKEN=<your_twitter_api_access_token>" >> .env # replace <your_twitter_api_access_token> with your Twitter API access token
echo "ACCESS_TOKEN_SECRET=<your_twitter_api_access_token_secret>" >> .env # replace <your_twitter_api_access_token_secret> with your Twitter API access 
python app.py

Locally

git clone https://github.com/mantreshkhurana/twitter-toxicity-detection-flask.git
cd twitter-toxicity-detection-flask
pip install -r requirements.txt
python setup.py # fill in the required details for Twitter API keys
python app.py

or

git clone https://github.com/mantreshkhurana/twitter-toxicity-detection-flask.git
cd twitter-toxicity-detection-flask
pip install -r requirements.txt
touch .env
echo "CONSUMER_KEY=<your_twitter_api_consumer_key>" >> .env # replace <your_twitter_api_consumer_key> with your Twitter API consumer key
echo "CONSUMER_SECRET=<your_twitter_api_consumer_secret>" >> .env # replace <your_twitter_api_consumer_secret> with your Twitter API consumer secret
echo "ACCESS_TOKEN=<your_twitter_api_access_token>" >> .env # replace <your_twitter_api_access_token> with your Twitter API access token
echo "ACCESS_TOKEN_SECRET=<your_twitter_api_access_token_secret>" >> .env # replace <your_twitter_api_access_token_secret> with your Twitter API access token secret
python app.py

or

git clone https://github.com/mantreshkhurana/twitter-toxicity-detection-flask.git
cd twitter-toxicity-detection-flask
pip3 install -r requirements.txt
touch .env
echo "CONSUMER_KEY=<your_twitter_api_consumer_key>" >> .env # replace <your_twitter_api_consumer_key> with your Twitter API consumer key
echo "CONSUMER_SECRET=<your_twitter_api_consumer_secret>" >> .env # replace <your_twitter_api_consumer_secret> with your Twitter API consumer secret
echo "ACCESS_TOKEN=<your_twitter_api_access_token>" >> .env # replace <your_twitter_api_access_token> with your Twitter API access token
echo "ACCESS_TOKEN_SECRET=<your_twitter_api_access_token_secret>" >> .env # replace <your_twitter_api_access_token_secret> with your Twitter API access token secret
python3 app.py

Navigate to http://127.0.0.1:5000/ in your web browser to use the app, else flask run -p 8000.

Usage

python app.py

If you want to run the app in a Window GUI:

python app.py --window
# or
python app.py -w

Use custom port:

python app.py --port 8000
# or
python app.py -p 8000

Features

  • Search for a Twitter user's recent tweets.
  • Dark/Light mode.
  • View a pie chart for profile's toxicity.
  • View user's profile picture, name, username, bio, location, website, following, followers, and tweet count*.
  • View images/videos in tweets.
  • View retweets and likes count for each tweet.
  • View the date and time of each tweet.
  • View the source of each tweet.
  • Twitter like feed.
  • Simple bot protection.
  • GUI Window Added.
  • Dialog to store API Keys in .env.
  • Images/Videos toxicity detection.

Project Structure

twitter-toxicity-detection-flask/
├── app.py
├── models/
│   └── hate_speech_model.csv
├── static/
│   ├── css/
│   │   └── style.css
│   └── js/
│       └── script.js
├── templates/
│   ├── index.html
│   ├── results.html
│   └── error.html
├── .env
├── .gitignore
├── README.md
└── requirements.txt

Contributing

Since this project took <12 hours to make you may find some bugs or you may want to add some features to it. You can contribute to this project by forking it and making a pull request(I am quite active on Github so if any issue arises I will try to fix it as soon as possible).

After forking:

git clone https://github.com/<your-username>/twitter-toxicity-detection-flask.git
cd twitter-toxicity-detection-flask
git checkout -b <your-branch-name>
# after adding your changes
git add .
git commit -m "your commit message"
git push origin <your-branch-name>

Credits

Author

About

This is a simple python program which uses a machine learning model to detect toxicity in tweets, developed in Flask.

Topics

Resources

License

Code of conduct

Security policy

Stars

Watchers

Forks

Releases

No releases published

Sponsor this project