Customers Churn

Predicting Customers Churning

Business Task: Telecom companies often struggle with retaining customers. I developed a predictive model to identify customers at risk of churning.

Approach:

• Data Collection: Gathered customer data including demographics, usage patterns, and service subscriptions.

• Exploratory Data Analysis (EDA): Conducted in-depth analysis to identify trends, correlations, and potential predictors of churn.

• Feature Engineering: Engineered new features such as customer tenure, usage intensity, and satisfaction scores.

• Model Selection: Evaluated various machine learning algorithms including logistic regression, random forest, and gradient boosting.

• Model Training and Evaluation: Trained models using a dataset split into training and testing sets. Evaluated performance metrics including accuracy, precision, recall, and F1-score.

Key Findings:

• Identified top predictors of churn: contract duration, monthly charges, and customer satisfaction.

• Developed actionable insights for reducing churn rates and improving customer retention strategies.

• Achieved an F1-score of 0.85 with the selected gradient boosting model.

Skills Utilized:

• Programming Languages: Python (NumPy, Pandas, Scikit-learn,  matplotlib, seaborn)

• Tools: Jupyter Notebook

• Machine Learning Techniques: These models (Logistic Regression, Random Forest, Gradient Boosting) were used. For this particular display, Random Forest was used.

Results: This predictive model enables telecom companies to proactively identify at-risk customers and implement targeted retention efforts, ultimately reducing churn rates and increasing customer lifetime value.