🧠 Inspiration

In many South Asian cities, civic sense often takes a backseat—streets littered with garbage, open potholes, stray animals, and neglect caused by lack of awareness. We wanted to build a system that makes citizen participation simple, rewarding, and AI-powered—bridging the gap between responsibility and technology. CivicSense was born from the idea that every smartphone can be an instrument of change.

💡 What it does

CivicSense is an AI-driven community platform where users upload photos of civic issues (like garbage, potholes, or waterlogging). The system automatically classifies the problem using a trained CNN model, assigns points based on severity, and updates a leaderboard of top contributors. Users can like, comment, and engage with others’ posts and tag their location (State → District) to help authorities and NGOs act faster.

🛠️ How we built it

Frontend: Streamlit for rapid prototyping and interactive UI. Backend (Prototype): Local JSON storage to simulate user and post data (no DB needed). AI Model: TensorFlow/Keras MobileNetV2 fine-tuned on 22 civic issue categories. Dataset: Scraped using Bing Image Downloader and manually curated images for key issues like potholes, stray animals, and plastic pollution. Location System: A custom-built India State–District mapping file for real geographic tagging.

⚡ Challenges we ran into

1.Collecting a balanced dataset for diverse civic issues was difficult due to overlapping visual contexts. 2.Designing an intuitive interface for non-tech-savvy users.

1. Implementing the idea in an interesting way so people actually take notice.

🏆 Accomplishments that we're proud of

Built a fully functional AI-powered civic engagement prototype end-to-end.

Implemented a working classification + points system that gamifies good citizenship.

Designed a clean, intuitive UI that encourages participation from people of all backgrounds.

Developed a modular structure ready for video upload and municipal data integration.

📚 What we learned

How AI and social good can coexist to build meaningful change.

The importance of human-centered design in civic applications.

Integrating AI inference in real-time user interfaces.

How gamification can drive behavior change in non-technical communities.

🚀 What’s Next for CivicSense — An AI-driven platform to empower citizens

🎥 Video Classification: Enable short video uploads so the AI can detect multiple civic issues in real-world motion. 🧠 Enhanced Model: Expand to 50+ issue categories using larger, more diverse datasets across South Asia. 🌏 Global Social Platform: Evolve CivicSense into a civic-focused social network—starting in South Asia, then scaling worldwide. 🏛️ Government Integration: Notify local authorities automatically when certain issue types are reported. 🤝 NGO Partnerships: Collaborate with NGOs to reward top contributors through recognition and incentive programs. 📢 Accessibility for All: Introduce WhatsApp and voice-based reporting to include non-tech-savvy citizens.

Built With

• bing-downloader
• python
• streamlit
• tensorflow