TweeToxicity

Twitter Profile Sentiment Analyzer

📖 Table of Contents

Table of Contents

• About The Project
• Screenshots
• Tech Stack
• Models
• Run Locally
• Environment Variables
• References

About the Project

TweeToxicity is a program that analyzes user profiles or hastags based on the recent tweets. The program utilizes machine learning to give Twitter users an appropriate score according to their tweets or retweets. This program is meant for educational purposes and no ill intetions existed prior to creating this program.

Screenshots

Tech Stack

Client:

• React
• Axios
• ChartJs

Server:

• NLTK
• Sklearn
• Tweepy
• Twitter API
• Flask Restful API
• BeautifulSoup

Models

DataSet Used for Training : Sentiment140

Model		Training		Testing
Name	Settings	Accuracy	F1 Score	Accuracy	F1 Score
Logistic Regression	Count Vect. & Lemmatization Used	79.63%	0.8008	78.71%	0.7929
	TF-IDF & Lemmatization Used	80.31%	0.8061	78.89%	0.7930
Multinomial Naive Bayes	Count Vect. & Lemmatization Used	78.48%	0.7838	78.71%	0.7756
	TF-IDF & Lemmatization Used	79.81%	0.7961	76.64%	0.7664
Bernoulli Naive Bayes	Count Vect. & Lemmatization Used	78.53%	0.7875	77.71%	0.7803
	TF-IDF & Lemmatization Used	80.15%	0.8052	77.68%	0.7791
Decision Tree Classifier	Count Vect. & Lemmatization Used, Decision Tree Parameters : {max_depth=50}	72.91%	0.7630	69.45%	0.7334
	TF-IDF & Lemmatization Used, Decision Tree Parameters : {max_depth=50}	73.66%	0.7667	69.13%	0.7297
Linear Support Vector Machine	Count Vect. & Lemmatization Used	82.92%	0.8309	78.38%	0.7867
	TF-IDF & Lemmatization Used	80.36%	0.8051	78.28%	0.7733
Random Forest Classifier	Count Vect. & Lemmatization Used, Random Forest Parameters : {max_depth=25}	74.65%	0.7615	74.04%	0.7566
	TF-IDF & Lemmatization Used, Random Forest Parameters : {max_depth=25}	74.80%	0.7619	74.00%	0.7553
Gradient Boosting Classifier	TF-IDF { min_df=5 } & Lemmatization Used . Gradient Boosting Parameters : {lr=1.25, n=100, depth=25}	74.80%	0.7619	74.00%	0.7553
	TF-IDF { min_df=5 } & Lemmatization Used . Gradient Boosting Parameters : {lr=1.25, n=100, depth=25}	85.99%	0.8626	77.49%	0.7791
XGBoost Classifier	Count Vect. & Lemmatization Used	75.29%	0.7661	75.21%	0.7662
	TF-IDF & Lemmatization Used	75.39%	0.7683	75.14%	0.7667

Run Locally

Clone the project

  git clone https://github.com/pri1311/TweeToxicity

Install dependencies in server folder.

  cd server
  python -m venv env
  source env/bin/activate
  pip install -r requirements.txt

Generate environment variables and fill in the values.

  cp .env.example .env

Your .env is ignored by git, which you can see in .gitignore, and so, it's safe!

Starting Development Server

  python server.py

Install dependencies in client folder.

  cd ../client # If you are in ./server
  npm i

Starting Client

  npm start

At the end of this, you should have

• server running at http://127.0.0.1:5002/
• new_client running at http://localhost:3000/

Environment Variables

To run this project, you will need to add the following environment variables to your .env file

API_KEY : Twitter API/Consumer Key

API_KEY_SECRET : Twitter API/Consumer Secret

BEARER_TOKEN : Twitter Bearer Token

ACCESS_TOKEN : Twitter Access Token

ACCESS_TOKEN_SECRET : Twitter Access Secret

References

• Pre-Processing :
  • Kaggle NLP 101 Preprocessing
• Machine-Learning-Projects Repo -> Twitter Sentiment Analysis
• Deploy a machine learning model using flask