DriveLedger

Description A Raspberry Pi-based system that retrieves simulated OBD (On-Board Diagnostics) data, classifies vehicle issues using AI, stores insights in Supabase, anchors proof-of-data on the Ethereum based Polygon Blockchain, and visualizes everything on a web dashboard.

Link to Live Website : https://drive-ledger.vercel.app/

Tech Stack

Table of Contents

1. Project Overview

2. Project Workflow Status

3. Tech Stack

4. Folder Structure

5. Setup and Installation

   • Prerequisites
   • Environment Variables
   • Docker Setup

6. Usage

---

Project Overview

This project demonstrates how to collect, analyze, and share meaningful vehicle OBD data without altering the car. It runs on a Raspberry Pi (or any Docker-supported machine), simulates OBD-II data, applies AI-based classification, records immutable proofs on a blockchain testnet, stores structured logs in Supabase, and provides a React dashboard for visualization.

---

Project Workflow Status

Component	Status	Notes
Simulated OBD Data Generator	✅ Complete	simulate_obd.py generates realistic OBD-like sensor data.
Supabase Schema + Integration	✅ Complete	Events and problems tables defined and data uploads working.
AI Classification Logic	✅ Complete	classify.py applies rule-based or trained model inference.
Dockerized Python Pipeline	✅ Complete	End-to-end pipeline runs in Docker, including Supabase upload.
Blockchain Hashing & Logging	✅ Complete	blockchain.py setup pending for submitting hash to Polygon testnet.
Frontend Dashboard	✅ Complete	Next js React + TypeScript frontend scaffolded, fetching + display pending.

---

Tech Stack

Layer	Technology	Purpose
Data Source	python-obd (simulated)	Generate realistic OBD data (RPM, temp)
AI	scikit-learn / TensorFlow Lite	Fault classification (e.g., overheating)
Blockchain	web3.py, Polygon testnet	Anchor hashed logs on-chain
Storage	Supabase (PostgreSQL)	Store event logs and AI results
Frontend	Next js React, TypeScript, Vercel	Dashboard for real-time insights
Containerization	Docker	Portable and reproducible deployment

---

Folder Structure

DriveLedger/
├── main.py                       # Master script: reads sensor data, runs ML, pins
│                                 # results to IPFS, logs to Supabase, mints NFT.
├── simulate_obd.py               # Generates / streams synthetic OBD-II readings.
├── supabase_client.py            # Small helper layer around the Supabase REST API.
│
├── generate_fake_training_data.py# Script to create synthetic datasets for model dev.
├── classify_tensorboard.py       # Visualise model-training metrics in TensorBoard.
├── export_all_to_csv.py          # Dump Supabase tables to CSV for offline analysis.
│
├── blockchain_commands.txt       # Handy on-chain CLI snippets & contract addresses.
├── training_data.csv             # Example raw training data (CSV format).
│
├── requirements.txt              # Python dependencies.
├── Dockerfile                    # Builds a Python + Node image so main.py can call
│                                 # `npx hardhat` inside the container.
│
├── package.json                  # Node dependencies (Hardhat, ethers, etc.).
├── package-lock.json             # Locked versions for reproducible Node builds.
│
├── used_ids.json                 # Local cache of already-minted NFT token IDs.
├── used_nonces.json              # (Optional) tracks contract nonces for advanced flows.
│
├── logs/                         # Trained ML artifacts.
│   └── …                         # e.g. model.tflite, scaler.pkl, fault_codes.pkl
│
├── obd_model_tf/                 # Original TensorFlow training notebooks/scripts.
│
├── driverledger-deploy/          # Hardhat project for Solidity contracts + mint script.
│   ├── scripts/mint.js           # Called by main.py to mint the NFT on Polygon.
│   └── …                         # Contracts, tests, Hardhat config, etc.
│
├── driveledgerwebsite/           # React/TypeScript front-end dashboard.
│   └── …                         # Components, pages, assets.
│
├── node_modules/                 # Installed automatically by npm / during Docker build.
└── __pycache__/                  # Python byte-code cache (auto-generated).

---

Setup and Installation

Prerequisites

• Docker & Docker Compose installed
• Python 3.9+
• Node.js & npm/yarn (for frontend)
• Supabase account & project
• Polygon testnet funds

Environment Variables

Create a .env file at project root with:

# Supabase
SUPABASE_URL=
SUPABASE_KEY=

# Blockchain (Polygon)
WEB3_PROVIDER_URL=
PRIVATE_KEY=

# Pinata IPFS
PINATA_API_KEY=
PINATA_SECRET_API_KEY=

Docker Setup

Build and run the service in Docker:

docker build -t driveledger .
docker run --rm driveledger

---

Usage

1. Clone the repo:

   git clone https://github.com/AbdulAaqib/DriveLedger.git
   cd DriveLedger

2. Populate .env.

3. Build and run with Docker (see above).

4. Access the frontend dashboard at http://localhost:3000 (or your Vercel URL).