EcoQuest

Inspiration

Our main inspiration for this project was the realization that lots of data for critical species are not representative of the actual population of the species on iNaturalist. Furthermore, people are driving everywhere and are not engaging in nature. EcoQuest was the perfect solution to both of those problems, specifically to enhance the quality of the current data for researchers

What it does

EcoQuest enables users to go on "quests" -- excursions with people who have similar personas to them. These quests send the user to find specific animals in San Diego, and then allow them to take pictures of those animals to send to iNaturalist to enhance the current data in the system

How we built it

Our front-end was a PWA, which allowed us to build an application that feels like a mobile app without specifically making it be iOS or Android native. Our system design for the back-end used multiple AWS services -- specifically we had multiple databases on DynamoDB and we also used Lambda services, S3, and configured IAM policies. We used 2 GPU, 2 core servers on JupyterHub for data processing and machine learning. We plan to deploy our machine learning models on AWS Sagemaker. We deployed this product on Vercel.

Challenges we ran into

Creating the machine learning models was incredibly challenging within the timeframe, and integrating the camera features -- both device and Ray-Ban, was difficult for us to be able to achieve. We also had multiple CORS issues in connecting the frontend and backend

Accomplishments that we're proud of

We are proud of building a very strong MVP that can be further developed into a full-fledged product that can be monetized

What we learned

Building a full-stack PWA with ML in 36 hours is chaotic, exhilarating, and completely worth it. Citizen science + gamification is a genuinely powerful combination for environmental impact. AWS Lambda + SageMaker made ML deployment feel approachable for the first time. Mock data is your best friend during a hackathon — ship the flow, wire the API later

What's next for EcoQuest

Phase 1: Expand species database, add push notifications, improve CV accuracy Phase 2: Native iOS & Android apps, wearable integrations beyond Ray-Ban, corporate team challenges Phase 3: Global city expansion, partner with conservation NGOs, API for researchers Adding revenue streams such as partnerships with local stores and educational institutions, researchers, and recreational organizations

Built With

• amazon-web-services
• computer-vision
• javascript
• jupyter-lab
• machine-learning
• pandas
• python
• vercel