Enron Email Spam Detection App

A web application that classifies emails as spam or ham, built using the Enron email dataset. The app demonstrates a full pipeline: data preprocessing, model training, and live deployment with Docker.

🔹 Live Demo

Try the frontend live here: https://enron-spam-frontend-1-0-0.onrender.com

Enter the email subject, message, and date

Click Classify Email to see if it’s spam or ham

🔹 Dataset

Original dataset: https://github.com/MWiechmann/enron_spam_data

License: GPL-3.0

This project uses a processed JSON version built from the original dataset for easier handling and faster model training

Both the original CSV and the converted JSON are included in the Docker image and container under the data folder

Project license: GPL-3.0

🔹 Model Training

Model trained in eda.ipynb using the original CSV dataset

Includes feature engineering, exploratory data analysis (EDA), and training of the classification model

Trained model is saved and used by the backend API for real-time predictions

Roughly 88% of the project code is in this notebook

ML Model Accuracy: 96%

🔹 Backend & Frontend Backend: FastAPI serving predictions at /predict

Frontend: Streamlit UI for classifying emails easily

Dockerized for consistent deployment

Deployed on Render for public access

🔹 Key Features

Preprocessed JSON dataset from Enron emails

Trained ML model integrated into backend API

Live demo via Streamlit frontend

Easy local testing using Docker Compose