WazeGrow: Predicting and Preventing User Churn for Waze

Using Data Science to Drive Retention and Growth

---

Project Overview

WazeGrow is a machine learning project designed to predict user churn for Waze, a leading navigation app. By identifying at-risk users and understanding churn drivers, Waze can implement proactive strategies to improve user retention and enhance user experience.

---

Table of Contents

• Project Goals
• Features
• Data
• Tools and Technologies
• Methodology
• Results
• Contributors
• License

---

Project Goals

1. Build a machine learning model to accurately predict monthly user churn.
2. Perform Exploratory Data Analysis (EDA) to uncover patterns and trends.
3. Provide actionable insights to Waze leadership for retention strategies.
4. Visualize data insights and model outcomes for technical and non-technical audiences.

---

Features

• EDA and Data Cleaning: Comprehensive analysis to identify trends and ensure data quality.
• Machine Learning Model: A predictive model with performance metrics like precision, recall, and F1-score.
• Visualizations: Tableau dashboards and Python plots for stakeholder communication.
• Actionable Recommendations: Data-driven insights for retention strategies.

---

Data

• The dataset includes anonymized Waze user data and features relevant to churn, such as usage patterns, demographics, and app engagement.
• Disclaimer: The dataset is synthetic and created for pedagogical purposes.

---

Tools and Technologies

• Programming Languages: Python (pandas, scikit-learn, matplotlib, seaborn)
• Version Control: Git and GitHub
• Visualization Tools: Tableau, Matplotlib, Seaborn
• Modeling: Logistic Regression, Random Forest, and other algorithms
• Project Management: PACE Framework

---

Methodology: PACE Framework

1. Plan: Drafted a project roadmap using the PACE framework to outline milestones and tasks.
2. Analyze: Conducted EDA and data cleaning to identify trends and correlations.
3. Construct: Built and validated machine learning models to predict churn.
4. Execute: Evaluated the model’s performance and created dashboards and visualizations in Tableau.

---

Expected Results

• Achieved 85% model accuracy in predicting churn.
• Identified key factors influencing churn, such as engagement frequency and app feature usage.
• Developed actionable strategies for user retention based on insights.
• Created interactive Tableau dashboards for non-technical stakeholders.

---

Contributors

• Jaime Orejarena: Lead Data Scientist and Repository Maintainer

---

Contact

For questions, suggestions, or collaboration, please reach out to Jaime Orejarena at orejarenajaime1979@gmail.com