SpotOn

Students waste time searching for study spaces. SpotOn shows live study space availability so you go to a free seat

Comment

Inspiration

Annoyed that I can’t find a room and I don’t have any time to study. Too often, time that should be spent working is wasted walking around libraries and buildings trying to find a free seat. The problem isn’t a lack of space—it’s a lack of visibility.

What it does

SpotOn is a real-time room availability and flow-tracking system. We track how many people enter and leave rooms through doorways, calculate current occupancy against capacity, and surface that information on a live website.

Students can:

  • See how many people are in each room and how many seats are available
  • Tell whether a room is getting busier or freeing up
  • View historical patterns like “Room A is usually free between 11:00–13:00”
  • Select a room and get notified when it becomes free around a chosen time
  • Discover “Hidden Gems” — quieter, underused spaces nearby

Staff can:

  • Monitor building-level and room-level occupancy
  • Identify bottlenecks and high-traffic areas
  • Improve layouts, signage, or timetables using real data
  • Know how many people remain in a building and where they are during evacuations

How we built it

Each doorway is fitted with a low-cost ultrasonic sensor device that detects the direction of travel, allowing us to distinguish entries from exits. These events are sent to a backend and stored centrally.

Architecture overview:

  • Edge devices (ESP32 + ultrasonic sensors): Count people entering and leaving rooms
  • MongoDB Atlas: Single source of truth for rooms, devices, and event data, with indexed time-based queries
  • AWS Lambda: Handles ingestion and API logic, including AI requests
  • Vultr Cloud Compute (2 vCPU, 8GB RAM): Always-on analytics engine for real-time congestion trends and rate-of-change calculations
  • Gemini API: Adds a natural-language interface so users can ask questions like “Where can I study that’s quiet right now?”
  • Web dashboard: Displays live occupancy, capacity, trends, patterns, and notifications

We designed schemas that link devices directly to rooms and ensure idempotent event handling, allowing the system to scale to hundreds or thousands of sensors without data corruption.

Challenges we ran into

  • Accurate direction detection: Getting reliable entry/exit counts with ultrasonic sensors required careful calibration and filtering
  • Real-time vs historical analytics: Balancing fast live updates with longer-term pattern analysis
  • Scalability: Designing data models and indexes that won’t fall apart as device counts grow
  • Latency: Ensuring room trends update quickly enough to be genuinely useful while people are on the move

Accomplishments that we’re proud of

  • Built a full end-to-end system: hardware → backend → analytics → AI → UI
  • Real-time occupancy and rate-of-change (“worth walking there or not”)
  • Practical emergency-response use case beyond just convenience
  • Low hardware cost with strong economies of scale
  • Clean, extensible schemas that can support future sensors (IR, cameras, etc.)

What we learned

  • Visibility solves more problems than capacity increases
  • Real-time data becomes far more valuable when paired with historical context
  • Hybrid cloud setups (serverless + always-on compute) are often the right answer
  • AI works best when grounded in structured, well-indexed data

What’s next for SpotOn

  • Pilot deployments across multiple buildings
  • Mobile-first UI and push notifications
  • Automatic floor-plan uploads and visual heatmaps
  • Smarter Gemini recommendations using user location and preferences
  • Expansion beyond universities to offices, events, and temporary venues

SpotOn turns wasted searching time into productive studying—and gives institutions the data they’ve never had about how their spaces are actually used.

Built With

Share this project:

Updates