cliffordnwanna/EXPRESSO_CHURN_PREDICTION_APP

The Expresso Churn Prediction App is a machine learning-based web application designed to predict customer churn for Expresso, a telecommunications service provider in Africa. This project demonstrates proficiency in data preprocessing, machine learning model training, and deploying interactive web applications using Streamlit. # Expresso Churn Prediction ## Overview This repository predicts customer churn for Expresso telecom customers and serves predictions through a Streamlit UI. The project has been revamped for portfolio quality with a focus on: - Reproducible training and evaluation - Leakage-safe preprocessing using a scikit-learn pipeline - Local trusted model artifact loading in the app - Clear metrics export for reporting ## Project Structure ``` EXPRESSO_CHURN_PREDICTION_APP/ ├── requirements.txt # Root-level pinned dependencies ├── training_code.py # Root-level shortcut to run training ├── APP/ │ ├── app.py # Streamlit inference app │ └── requirements.txt # App-level pinned dependencies ├── DATA/ │ ├── dataset # Contains dataset download link │ └── Expresso_churn_dataset.csv # Raw dataset (download separately) ├── MODEL/ │ ├── churn_model_bundle.joblib # Generated by training script │ └── metrics.json # Generated by training script ├── SCRIPTS/ │ ├── data_cleaning.py # Optional data cleaning step │ └── training_code.py # Full training pipeline └── README.md ``` ## Dataset **Important:** Before running the training pipeline, you must download the Expresso dataset. 1. Download the dataset from this link: Expresso Churn Dataset (Google Drive) 2. Save the downloaded file as: ``` DATA/Expresso_churn_dataset.csv ``` 3. Verify the file exists in the DATA folder before running the training script. ## Installation ### 1. Download Dataset First (Required) Before installing or running anything, **download the dataset**: - Go to: Expresso Churn Dataset (Google Drive) - Save the file as: `DATA/Expresso_churn_dataset.csv` ### 2. Install Dependencies From the repository root, choose one: **Option A (Recommended):** ```bash pip install -r requirements.txt ``` **Option B (Alternative):** ```bash pip install -r APP/requirements.txt ``` …