This is a complete end-to-end machine learning application that predicts whether bank customers will leave (churn) using logistic regression. It transforms a traditional data science workflow into an interactive web application that anyone can use through their browser.
This is a complete end-to-end machine learning application that predicts whether bank customers will leave (churn) using logistic regression. It transforms a traditional data science workflow into an interactive web application that anyone can use through their browser.
- 🤖 Machine Learning Model Algorithm: Logistic Regression (a classification algorithm)
Purpose: Predicts if a customer will stay (0) or leave (1) the bank
Input: Customer data (credit score, age, balance, tenure, etc.)
Output: Churn probability (0-100%) and binary prediction
- 🌐 Live Web Application Interactive Interface: Users can input customer details through forms
Real-time Predictions: Instant churn probability calculations
No Coding Required: Business users can make predictions without technical knowledge
- 📊 Analytics Dashboard Model Performance Visualization: ROC curves, confusion matrices
Feature Importance: Shows which factors most influence churn
Classification Reports: Detailed accuracy metrics
Data Upload → Users upload customer data (Excel/CSV)
Automatic Preprocessing → Cleans and prepares data for modeling
Model Training → Trains logistic regression on the data
Live Predictions → Users can test new customer data
Performance Analysis → Visualizes how well the model performs
https://bankchurnanalytics.streamlit.app/