🚗 GetAround — Delay Analysis & Pricing Prediction

Certification CDSD — Data Science & Deployment Project — Jedha Bootcamp

---

📌 Project Overview

GetAround is a peer-to-peer car rental platform. Late vehicle returns create friction for subsequent rentals, leading to customer dissatisfaction and cancellations.

This project addresses two strategic challenges:

• Operational optimization — Analyzing late checkouts and simulating minimum delay thresholds to reduce conflicts between consecutive rentals.
• Pricing optimization — Serving a Machine Learning model via a production API to help owners set optimal daily rental prices.

---

🔗 Production Links

Service	URL
📊 Dashboard	https://huggingface.co/spaces/Dreipfelt/getaround-dashboard
🔌 API	https://Dreipfelt-getaround-api.hf.space
📄 API Docs	https://Dreipfelt-getaround-api.hf.space/docs
💻 GitHub	https://github.com/Data-Science-Designer-and-Developer/Project_GetAround

---

🎯 Business Objectives

Delay Management

• Measure how often drivers return cars late
• Quantify the impact on subsequent rentals
• Simulate different minimum delay thresholds (0 to 720 minutes)
• Help Product Management choose:
  • an optimal delay threshold
  • an appropriate scope (all cars vs Connect only)

Pricing Optimization

• Train a ML model on car characteristics
• Serve predictions via a REST API
• Allow real-time price prediction through a /predict endpoint

---

📊 Dashboard

The interactive dashboard allows Product Managers to:

• Visualize the distribution of late checkouts
• Compare Connect vs Mobile check-in types
• Simulate the trade-off between blocked rentals and resolved issues
• Filter by scope and threshold in real time

🔗 https://huggingface.co/spaces/Dreipfelt/getaround-dashboard

---

🤖 Machine Learning API

Model

Property	Value
Algorithm	Random Forest Regressor
Target	rental_price_per_day (€)
R² score	~0.68
Features	28 (mileage, engine_power, fuel, color, car_type, options...)

Endpoint /predict

• Method : POST
• Input : JSON with key input — list of lists

curl -X POST "https://Dreipfelt-getaround-api.hf.space/predict" \
     -H "Content-Type: application/json" \
     -d '{"input": [[150000, 120, 1, 1, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0]]}'

Response :

{"prediction": [104.75]}

📄 Full documentation : https://Dreipfelt-getaround-api.hf.space/docs

---

🗂️ Repository Structure

Project_GetAround/
├── api/                        # FastAPI application
│   ├── app.py                  # API endpoints
│   ├── Dockerfile              # Docker configuration
│   └── feature_names.json      # Model feature names
│
├── dashboard/                  # Streamlit dashboard
│   ├── app.py                  # Dashboard application
│   └── requirements.txt
│
├── notebooks/                  # Jupyter notebooks
│   ├── 01_EDA_delays.ipynb     # Delay analysis
│   └── 02_ML_pricing.ipynb     # ML model training
│
├── .gitignore
└── README.md

---

🛠️ Tech Stack

Category	Tools
Language	Python 3.10
Dashboard	Streamlit, Plotly
API	FastAPI, Uvicorn
ML	Scikit-learn, Random Forest
Deployment	Hugging Face Spaces, Docker
Version Control	Git, GitHub

---

⚙️ Local Setup

# Clone the repo
git clone https://github.com/Data-Science-Designer-and-Developer/Project_GetAround.git
cd Project_GetAround

# Install dependencies
pip install -r dashboard/requirements.txt

# Run the dashboard
streamlit run dashboard/app.py

# Run the API
cd api
uvicorn app:app --reload
# API available at http://localhost:8000

---

👤 Author

Frédéric CDSD Candidate — Data Scientist Jedha Bootcamp