Social Media Bot Detection and Classification

This repository collects the notebooks, model weights, and notes for a social media classification project built around tweet text, user descriptions, and metadata features. The workflow appears to combine LSTM-based text modeling with additional structured features for classifying social accounts / tweets.

What is in this repo

• notebooks/combined_tweet_model.ipynb - combined tweet pipeline
• notebooks/social_tweets_model.ipynb - tweet-level pipeline
• notebooks/social_users_model.ipynb - user-description pipeline
• notebooks/exploratory/ - additional experimentation notebooks
• models/ - saved TensorFlow/Keras weights
• docs/notes/ - supporting notes and feature lists
• docs/figures/output_format.png - small reference for model output formatting

Main idea

The code in the notebooks uses preprocessing for tweet text and user text, plus metadata-derived features, to train classification models. The saved models suggest a TensorFlow/Keras workflow with LSTM-based sequence modeling.

Important cleanup notes

• Several notebooks use hard-coded Windows paths like D:\RPP\...; those paths need to be replaced with paths relative to your own machine.
• The raw datasets are not included here.
• Some filenames were cleaned up to make the repo easier to read and navigate.

Suggested setup

Install the usual Python stack used in the notebooks:

pip install -r requirements.txt

Then open the notebooks in Jupyter and update the local data paths before running them.

Folder structure

notebooks/
models/
docs/
src/

GitHub push flow

This folder is ready to be turned into a Git repo and pushed to your account.

git init
git add .
git commit -m "Organize social media bot detection project"
git branch -M main
git remote add origin <your-repo-url>
git push -u origin main