regmul/Streamlit_checkpoint1-Expresso_Churn_Prediction

This project uses the Expresso churn dataset from Zindi to predict customer churn for a telecom company in Mauritania and Senegal. The workflow includes data cleaning, feature engineering, and training a Logistic Regression model, deployed in a Streamlit app for interactive predictions. # **Expresso Churn Prediction – Logistic Regression & Streamlit App** This project is based on the Expresso Churn Prediction Challenge originally hosted on the Zindi platform. The goal is to build a machine learning pipeline to predict customer churn for Expresso, a telecommunications company operating in Mauritania and Senegal. Customer churn prediction is a crucial task in the telecom industry since it helps identify customers likely to leave, enabling proactive retention strategies. **📁 Dataset** The dataset contains information on 2.5M clients with more than 15 behavioral and usage features. It is not included in this repository due to its large size. 🔗 You can download the dataset from the challenge: 👉 Expresso Churn Dataset (Zindi) Once downloaded, place it in a folder named data/ (ignored by Git). **⚙️ Workflow** **Data Exploration & Cleaning** Checked dataset info, missing values, duplicates, and outliers. Generated a Pandas Profiling Report for insights. Cleaned and prepared the dataset for modeling. **Feature Engineering** Encoded categorical features. Normalized continuous features. Selected relevant predictors for churn modeling. **Model Training** Trained a Logistic Regression model to predict churn. Evaluated performance on test data. Saved the model using pickle for deployment. **Deployment with Streamlit** Built an interactive Streamlit app for predictions. Users input customer details The app predicts churn likelihood in real time. You can try the **live Streamlit app** here:([expresochurnprediction.stre…])