EcoNova

Inspiration

Climate change is overwhelming, but nobody actually knows how their daily choices add up. We wanted to make carbon footprints personal, measurable, and competitive. If Strava can make running addictive, why can't we do the same for sustainability?

What it does

Snap a photo of any receipt — Gemini reads every line item, identifies the product, and estimates its CO₂ using IPCC emission factors.
Paste your bank statement and Gemini classifies every merchant automatically — Shell becomes fuel, Uber becomes rideshare, H&M becomes fast fashion — each with a CO₂ score.
Real-time GPS commute tracking detects your transport mode by speed — walking, cycling, bus, car, or train — no manual input, just open the app and move.
Every emission gets categorised into food, transport, energy, shopping, and digital — with a full breakdown dashboard so you can see exactly where your footprint comes from.
Live leaderboard ranks you against friends with the lowest score winning — updated in real time as everyone logs.
Compare your footprint against your national average and your global percentile — see exactly where you stand against the average Canadian.

How we built it

FastAPI backend with PostgreSQL for data storage. Gemini 2.0 Flash powers all the parsing — we use its vision capability to read receipt images directly, and its JSON mode to extract structured CO₂ data from bank transactions. Emission factors are sourced from IPCC, DEFRA 2024, and Our World in Data. The GPS tracker uses the browser Geolocation API with speed thresholds to classify transport mode without any manual input.

Challenges we ran into

Honestly, the biggest challenge was time. With only 8 hours to go from idea to working backend is brutal. We had to make fast decisions about what to cut, what to keep, and what to duct-tape together. Python 3.13 on Windows had no prebuilt wheel for asyncpg so we had to swap the database driver mid-build.

Accomplishments that we're proud of

Gemini reading a receipt photo and returning accurate CO₂ estimates with confidence scores in under 3 seconds. The bank transaction parser correctly identifying Shell as fuel, Uber as rideshare, and H&M as fast fashion purely from merchant names and spend amounts. Getting the full stack running end to end — image upload to CO₂ score to leaderboard — in a single hackathon.

What's next for EcoNova

Weekly AI-generated report from Gemini — a plain-English narrative of your week, what drove your emissions up or down, and one specific challenge for next week.
Personalised coaching tips from Gemini — not generic advice, but specific swaps based on your actual breakdown ("swap 2 beef meals for chicken this week = −11 kg CO₂")
Carbon Q&A chatbot powered by Gemini — ask anything and it answers using your real footprint data as context.
Daily and weekly challenges — go meatless, skip the car, drop the thermostat — each with a verified CO₂ reduction when completed.
Streak tracking for consecutive low-carbon days — keeps the competition going between big logging events.
Social activity feed showing what everyone in your group is doing — "Jordan took a flight ✈ +650 kg", "Alex went vegan today 🌱 −4.5 kg"
Connecting directly to bank APIs via Plaid for automatic transaction sync.
A mobile app with always-on GPS tracking. Group challenges — offices, universities, friend groups competing weekly.
And a carbon offset marketplace where your leaderboard rank unlocks verified offset projects you can actually fund.