Waze User Churn Prediction

Project Overview

The Waze User Churn Prediction project aims to predict user churn for the Waze application, leveraging advanced data science techniques. This project involves analyzing user data to identify patterns and build predictive models that can forecast whether users are likely to stop using the application. The goal is to provide actionable insights to help the company enhance user retention strategies.

Project Objectives

• Predict User Churn: Develop models to predict user churn using historical data.
• Feature Engineering: Identify and engineer features that are significant predictors of churn.
• Model Evaluation: Assess model performance using various metrics and validate its effectiveness.

Key Deliverables

1. One-Page Summary: A summary document highlighting key findings, insights, and recommendations based on the analysis.
2. Complete Code Notebook: A Jupyter Notebook with the complete code, including data processing, feature engineering, model training, evaluation, and results.

Project Stages

1. Plan: Defined the problem, objectives, and steps required to achieve the goals.
2. Analyze: Performed Exploratory Data Analysis (EDA) to understand the dataset and uncover patterns.
3. Construct: Built predictive models using Logistic Regression and tree-based models (Decision Tree and Random Forest). Conducted feature engineering and model tuning.
4. Execute: Evaluated model performance and interpreted the results to provide actionable insights for user retention strategies.

Technologies and Tools

• Programming Languages: Python
• Libraries and Frameworks: Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn
• Tools: Jupyter Notebooks

Models Used

• Logistic Regression: Applied to predict user churn based on historical data.
• Tree-Based Models:
  • Decision Tree: Used for initial modeling.
  • Random Forest: Employed in two rounds—before and after feature engineering—to improve accuracy and robustness.

Evaluation Metrics

• Accuracy: Measures the overall correctness of the model.
• Precision and Recall: Assesses the model’s ability to correctly identify churned users.
• F1 Score: Provides a balance between precision and recall.

Data Visualization

Included visualizations to illustrate key trends, feature importance, and model performance, enhancing the interpretability of the results.

Ethical Considerations

Ensured that the analysis and predictions were conducted with respect to user privacy and data security. The insights were used responsibly to support business decisions aimed at improving user experience.

• Email: loviaeb@gmail.com
• LinkedIn: linkedin.com/in/lovia-edassery

---

Feel free to adjust any sections based on the specific details and structure of your project.