Churn Prediction Using Random Forest Classifier

Developed a Streamlit app to predict customer churn for PowerCo, a gas and electricity supplier. You can access the app using this link.

Overview

This project uses a Random Forest Classifier to predict customer churn based on various factors such as electricity, gas, and power consumption. The model also considers net margin evolution, price changes, and sensitivity during off-peak, peak, and mid-peak periods. The complete process is detailed in the application, from data loading and exploratory data analysis to feature engineering, model building, and predictions.

Features

• Data Loading: Import and prepare datasets for analysis.
• Exploratory Data Analysis (EDA): Analyze data to understand patterns and relationships.
• Feature Engineering: Create and select relevant features for the model.
• Model Building: Train a Random Forest Classifier on the prepared data.
• Feature Importance: Extract and visualize important features used by the model.
• Prediction: Make predictions based on the trained model.

Important Features

Tools

Python, Streamlit, Machine Learning