Building a Customer Churn Prediction Model with Logistic Regression

Description and use

A logistic regression model was built to predict telecom customer churn, achieving ~80% accuracy.

Problem

Identify telecom customers likely to churn to enable proactive retention strategies.

Outcome

Achieved ~80% accuracy on validation set, identified contract type as most informative categorical feature, and demonstrated end‑to‑end ML workflow from data prep to model interpretation.

Workflow

• Explore dataset: target variable, missing values, feature distributions
• Compute feature importance: risk ratio/difference, mutual information, correlation
• Encode categorical features with Scikit‑Learn DictVectorizer to 45‑column matrix
• Train logistic regression model on encoded data
• Evaluate model accuracy (~80%) and interpret weights
• Create smaller model with selected features and manual prediction calculation

Evidence

I trained a logistic regression model using Scikit‑Learn and achieved approximately 80% accuracy on the validation set.

Source

One finding that stood out was that contract type had the highest mutual information among the categorical features we examined.

Source