🏛️ CivicProof: AI-Driven Micro-Accountability Engine

🔴 View Live Demo Here (Note: Replace with your actual deployment link when ready)

Bridging the Trust Deficit between Government Promises and Ground Reality.

CivicProof is an AI-powered governance platform designed to automate the verification of hyper-local civic works (like streetlight repairs and pothole filling). It forces transparency by requiring field workers to upload visual proof, which is then audited by a Computer Vision AI before marking a task as "Complete" and automatically notifying local citizens.

---

🚀 See It In Action

Real-Time Geospatial Monitoring

Our dynamic Streamlit dashboard tracking booth-wise activity and task verification in real-time.

The AI Engine Under the Hood

FastAPI backend instantly running YOLOv8 and OpenCV logic to verify uploaded evidence.

---

🛑 The Problem: "Ghost Work"

Local municipalities spend millions on last-mile infrastructure but face a massive "Trust Deficit."

• 📉 Lack of Verification: Authorities cannot physically inspect thousands of daily micro-tasks.
• 👻 Fake Reporting: Contractors often mark jobs as "Complete" without actually doing the work.
• 🔇 Citizen Disconnect: Residents see broken infrastructure, report it, but never receive transparent updates.

💡 The Solution: Visual Proof Protocol

CivicProof replaces manual checks with an Automated Trust Loop:

1. Assign 📍: A task (e.g., "Fix Streetlight") is assigned to a local booth.
2. Execute & Upload 📸: The worker finishes the job and uploads a photo to the portal.
3. AI Audit 🧠: The YOLOv8 + OpenCV engine analyzes the image for context (is it a street?) and execution (is the light actually glowing?).
4. Notify 📲: Once verified, the system automatically sends an SMS notification to the specific voters registered in that booth area.

---

✨ Key Features

• 🔍 Smart Glare Detection (Computer Vision): Custom blob-detection algorithms distinguish between a working lightbulb and a bright daytime sky, preventing "false positive" verifications.
• 🗺️ Geospatial War Room: A real-time dashboard providing authorities with a heat map of pending vs. verified complaints.
• 🔄 Automated Citizen Loop: Closes the feedback loop by linking verified tasks directly to the voter registry database for localized updates.
• 🛡️ Fraud Prevention: Automatically rejects dark, blurry, or context-inaccurate photos.

---

📸 System Previews

1. The Command Center Dashboard

Real-time metrics, geospatial booth mapping, and task tracking.

2. Live Activity Map

Monitoring task statuses (Pending, Verified, Rejected) across different city zones.

---

🛠️ Technology Stack

• Backend: Python, FastAPI, Uvicorn
• AI/Machine Learning: Ultralytics YOLOv8 (Object Detection), OpenCV (Luminosity/Blob Analysis), NumPy
• Database: SQLite & SQLAlchemy (Knowledge Graph modeling of Voters ↔ Booths ↔ Tasks)
• Frontend: Streamlit, Pandas

---

💻 Quick Start Guide (Local Deployment)

Prerequisites

• Python 3.9+
• Git

Installation Steps

1. Clone the repository:

   git clone [https://github.com/yourusername/CivicProof.git](https://github.com/yourusername/CivicProof.git)
   cd CivicProof

2. Install dependencies:

   pip install -r requirements.txt

3. Initialize and Seed the Database: This creates a local SQLite database and populates it with dummy booths, voters, and tasks.

   python seed.py

4. Start the FastAPI Backend:

   uvicorn main:app --reload

   The API will be live at http://127.0.0.1:8000 (Access the Swagger UI at /docs).

5. Launch the Streamlit Dashboard: Open a new terminal window and run:

   streamlit run dashboard.py

   The dashboard will automatically open in your browser.

---

🤝 Open for Collaboration

This project is open-source and looking for contributors! Areas where we'd love your help:

• 🎨 Frontend Polish: Enhancing the Streamlit UI with custom CSS and better Map integrations (Folium).
• 🧠 AI Expansion: Training custom YOLO models to detect specific civic issues like garbage dumps, potholes, and waterlogging.
• 🔐 Security: Implementing EXIF GPS data extraction to ensure uploaded photos match the exact task coordinates.

Feel free to Fork the repository and submit a Pull Request!

---

Developed for robust, transparent, and hyper-local governance.

---

Essential Steps for a Professional Manual Upload

To make sure this renders perfectly on GitHub, double-check these steps when uploading:

1. Root Directory Placement: Ensure your README.md, main.py, dashboard.py, ai_engine.py, seed.py, and requirements.txt are all sitting directly in the main folder of your repository.
2. Image Placement: Because of how the markdown is written above, the images and GIFs (Screenshot 2026-02-16 134807.png, EditCivicPGIF.gif, etc.) must be uploaded to the exact same main folder as the README. If you put them in an "images" folder, the links will break unless you update the markdown to say images/EditCivicPGIF.gif.
3. The .gitignore File: Manually create a file named exactly .gitignore (with the dot) in your GitHub repo and add *.db and __pycache__/ to it. This shows recruiters you know how to keep repositories clean of unnecessary system files.
4. Update the URL: Don't forget to replace https://your-live-url-goes-here.com at the top of the README once you eventually host it online (like on Render or Streamlit Community Cloud).