This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
I built EcoPulse — AI Bio-Blitz, an AI-powered biodiversity exploration application designed to encourage people to spend more time outdoors and discover the biodiversity around them.
The idea is simple:
Use AI as a reason to go outside, not as a reason to stay on the screen.
EcoPulse allows users to select an exploration environment, difficulty, and type of biodiversity they want to discover. Based on these choices, the application generates five outdoor biodiversity missions.
For example:
- 🌱 Find a plant with an interesting shape and observe its leaves.
- 🐝 Look for an insect and observe what it is doing.
- 🌿 Find two different plants and compare their appearance.
- 🌸 Look for a small sign of biodiversity that you might normally miss.
Users can then go outdoors, discover something interesting, and upload a photograph of it.
The AI analyzes the image and provides:
- 🔎 Possible identification
- 📊 Confidence level
- 📖 Beginner-friendly explanation
- 🌿 Biodiversity information
EcoPulse also connects the identified organism to the iNaturalist API for additional biodiversity and taxonomic information.
Users can complete missions and earn BioPoints.
The EcoPulse Loop
Choose exploration preferences
↓
AI generates 5 missions
↓
Go outside 🌱
↓
Find something interesting
↓
Take a photo 📸
↓
AI analyzes the discovery
↓
Possible identification
↓
Biodiversity information
↓
Complete a mission
↓
Earn BioPoints
↓
Keep exploring
Demo
🎥 Watch the EcoPulse Demo:
The demo shows the main working flow:
- Selecting an exploration environment
- Selecting the difficulty
- Selecting the biodiversity type
- Generating five AI missions
- Uploading a discovery photo
- Analyzing the discovery
- Showing the AI identification and confidence
- Retrieving biodiversity information
Code
💻 GitHub Repository:
https://github.com/meenakshi930/ecopulse-ai
How I Built It
I built EcoPulse using Python and Streamlit, with separate components for mission generation, image analysis, and biodiversity data retrieval.
AI Mission Generation
The user selects an environment, difficulty, and biodiversity type. These inputs are used to generate five outdoor biodiversity missions.
The application validates the generated response and includes fallback missions in case of an AI or API failure.
Vision-Based Discovery
Users can upload a photograph of something they discover outdoors.
The vision model analyzes the image and provides:
- Possible identification
- Confidence level
- Beginner-friendly explanation
- Biodiversity information
The response is returned in a structured format so that the application can validate and display it reliably.
Biodiversity Information
After getting a possible identification, EcoPulse queries the iNaturalist API to retrieve:
- Scientific name
- Common name
- Taxonomic rank
- Observation count
This separates AI-based visual interpretation from external biodiversity data.
Mission Progress
EcoPulse uses Streamlit session state to keep track of:
- Generated missions
- Completed missions
- BioPoints
- Discovery results
- Biodiversity information
Completing a mission awards BioPoints.
Technology Stack
| Technology | Purpose |
|---|---|
| Python | Main programming language |
| Streamlit | Web application interface |
| AI models | Mission generation and image analysis |
| iNaturalist API | Biodiversity information |
| Pillow | Image processing |
| Pydantic | Structured AI response validation |
| Requests | API communication |
| python-dotenv | Environment variable management |
Why Does Open Innovation Matter?
Open innovation makes it possible for students and independent developers to experiment with AI and build applications around real-world problems.
For EcoPulse, AI is not simply being used as a chatbot. It is part of an outdoor exploration experience.
The project combines AI-based visual understanding with biodiversity data to encourage people to observe and learn about nature around them.
Most importantly, EcoPulse tries to reverse the usual relationship between technology and screen time.
Instead of encouraging people to stay inside and keep scrolling, the application uses technology to encourage them to step outside.
The phone becomes a tool for exploration, not the destination.
What I Learned
Building EcoPulse helped me learn about:
- Integrating AI into a Python application
- Prompt design
- Vision AI
- Structured AI responses
- JSON and schema validation
- REST API integration
- Working with external biodiversity data
- Streamlit session state
- Handling AI and API failures
- Building an end-to-end AI application
One important lesson was that AI output should not always be treated as ordinary text.
AI responses are probabilistic, so application logic should validate important outputs before using them.
What I Would Improve Next
EcoPulse is currently an MVP.
One important future improvement would be Mission-to-Photo Verification.
For example:
Mission:
"Find a flowering plant."
↓
User uploads a photo
↓
Vision AI analyzes it
↓
Mission verified
↓
BioPoints awarded
This would make the connection between the outdoor mission and the AI discovery more meaningful.
Future versions could also explore better species identification, more biodiversity categories, location-aware missions, richer exploration history, and automatic mission verification.
My Agent Session
I did not use a DevRelay agent session for this project.
Prize Categories
I am submitting EcoPulse for the Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass.
Top comments (0)