Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🚗 Predicting Uber Driver Activation

📌 Overview

This project analyzes a real-world business problem: understanding what factors influence whether someone who signs up for Uber actually goes on to complete their first trip as a driver. By leveraging predictive modeling and data visualization, we aim to help Uber refine its driver onboarding strategy and improve activation rates.


💡 Inspiration

When brainstorming project ideas, we wanted something enjoyable and impactful. Our goal was to extract valuable insights from a relatively small dataset—showcasing how companies can still make smart decisions with limited data. We chose a StrataScratch dataset because it offered an authentic, real-world scenario where we could dig deep into a company's operational data and draw meaningful conclusions.


🔍 Problem Statement

How can Uber use limited signup data to predict whether a new driver will complete their first trip?
Our project dives into identifying the most significant predictors of driver activation, offering actionable recommendations to improve recruitment and onboarding processes.


⚙️ How It Works

  • Dataset: Contains driver signup data with features such as signup channel, device, referral source, and more.
  • Target Variable: started_driving (binary classification: 1 if the driver completed their first trip, 0 otherwise)
  • Models Used:
    • Random Forest Classifier
    • XGBoost (LightBoost)
  • Evaluation Metrics:
    • Accuracy
    • Feature Importance
    • Confusion Matrix

🔨 Built With

  • 🐍 Python (pandas, scikit-learn, xgboost, matplotlib, seaborn)
  • 📊 Power BI for dashboards and data storytelling
  • 🤖 Machine Learning for classification and feature selection

📈 Key Features & Process

✅ Predictive Modeling

🔍 Exploratory Data Analysis (EDA)

🧹 Data Preprocessing & Feature Engineering


🚧 Challenges

  • Dealing with missing or inconsistent values.
  • Balancing between adding engineered features and preventing overfitting.
  • Translating model insights into actionable business recommendations.

🏆 Accomplishments

  • Created an end-to-end machine learning pipeline.
  • Built an interactive Power BI dashboard to support business storytelling.
  • Collaborated effectively as a team to deliver both technical depth and business value.

🎓 Learning Outcomes

  • Applied ML algorithms to a real business classification problem.
  • Gained experience in building interpretable models and refining them through feature selection.
  • Developed compelling visualizations in Power BI that drive decision-making.
  • Learned how to turn raw data into a coherent and persuasive story for stakeholders.

🔮 What’s Next for the Unforeseen

If external datasets become available, the next logical step is to compare our findings with a competitor like Lyft. This could uncover platform-specific behavioral insights and highlight where Uber might improve to gain a competitive edge in driver acquisition and retention.


📂 Folder Structure

├── data/                    # Raw and cleaned datasets
├── notebooks/              # Jupyter Notebooks with EDA and modeling
├── visuals/                # Power BI exports and graphs
├── models/                 # Trained model files (if applicable)
└── README.md               # Project documentation (you are here)

🤝 Team

Project by: Kento Morita, Phuong Duong, Vincent Perez, Ariel Liang We’re a group of business analytics students passionate about solving real-world problems through data and machine learning.

About

We've tried building prediction model to help understand what factors are best at predicting whether a signup will start to drive, and offer suggestions to operationalize those insights to help Uber.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages