Wi-Fi Sense

🚀 Inspiration

Every building today is filled with Wi-Fi signals—but most of that data goes unused. We were inspired by the idea of transforming existing infrastructure into something smarter and greener. What if we could estimate occupancy and optimize building operations—using nothing more than the Wi-Fi that's already there?

💡 What it does

WiFi Sense passively ingests Wi-Fi device counts from access points, links them to physical zones, and estimates real-time or historical occupancy using machine learning. The result is a zone-level heatmap dashboard that helps optimize energy usage, custodial services, and space management.

🛠 How we built it

Frontend: Built in Next.js and Konva.js for floorplan heatmap rendering

* Backend: Python-based daily data bridge script + ML model (Linear Regression) in progress

⚔️ Challenges we ran into

Mapping APs to physical zones was harder than expected
Creating a frontend that feels live, even with daily batch data
Scaling prediction models across different types of zones

🏆 Accomplishments that we're proud of

Built a real-time-feeling dashboard in under 12 hours
Created a cost-effective, privacy-first alternative to camera-based sensing

📚 What we learned

Existing infrastructure is gold—if you know how to mine it
Frontend interactivity (like map overlays) is just as important as backend accuracy
Designing with privacy in mind from Day 1 is totally doable—and marketable