🚚 Delivery ETA Prediction System

A full-stack project for predicting delivery estimated time of arrival (ETA) using machine learning and Flask API, with PostgreSQL for data storage and a simple frontend for interaction.

---

🔧 Features

• 🤖 ML Model (Linear Regression) to predict delivery ETA based on distance and time features
• 🛠️ Flask API for ETA prediction and logging delivery data
• 🗄️ PostgreSQL database to store ETA logs including driver and order info
• 🌐 Simple frontend HTML pages for submitting delivery requests and viewing logs
• 📊 Integration with Metabase for analytics and dashboarding

---

🧰 Tech Stack

Component	Technology
ML Model	Python, scikit-learn
Backend API	Python, Flask
Database	PostgreSQL
Frontend	HTML, JavaScript
Environment Config	python-dotenv
Analytics	Metabase

---

🗃️ Folder Structure

delivery-eta-prediction/
├── model/               # Saved ML model files (e.g. delivery_eta_lr.pkl)
├── templates/           # Static files for frontend (HTML, CSS, JS)
├── mock_data.py         # Generating mock data
├── data/                # Saved mock data files
├── model.ipynb          # Training models
├── app.py               # Flask API and server
├── utils.py             # Utility functions (distance calculation, ETA prediction)
├── .env                 # Environment variables (DB credentials, etc.)
└── README.md            # Documentation (this file)

🚀 How It Works

Build and train ML model

• Use mock delivery data with features: distance (km), hour of day, and weekday
• Train a Linear Regression model to predict ETA in minutes
• Save the trained model with joblib

Flask API Endpoints

• /predict_eta (POST):
  Accepts JSON with driver info and coordinates, returns predicted ETA and status message, saves log entry to the database
• /eta_logs (GET):
  Returns all logged ETA predictions with timestamps

Frontend HTML pages

• ETA form for entering delivery info and receiving ETA predictions
• ETA logs page to view historical prediction records
• Optional dashboard page for visualization or embedding Metabase dashboards

Database (PostgreSQL)

• Stores delivery logs with driver_id, order_id, coordinates, ETA, status message, and timestamp

---

📊 Sample Data Structure

Column Name	Description
order_id	Unique identifier for each order
driver_id	Delivery driver's ID
current_lat	Latitude of current location
current_lng	Longitude of current location
dropoff_lat	Latitude of dropoff location
dropoff_lng	Longitude of dropoff location
eta_minutes	Predicted estimated time in minutes
status_message	ETA status ("Arrived", "En route", etc.)
ts	Timestamp of prediction/log entry

---

✅ Deliverables

• Mock data creation and model training (Linear Regression)
• Utility functions for haversine distance calculation and ETA prediction
• Flask API for prediction and data logging
• Frontend HTML pages with input form and logs table
• PostgreSQL database integration
• Environment variables handling with .env file
• Basic Metabase dashboards for delivery analytics (optional)

---

📜 License

MIT License — Free to use, modify, and extend for learning or portfolio projects.

---

👩‍💻 Author

Angela Loro
Developed during internship at LAMINA Studios
GitHub: github.com/llaight