Inspiration
Lack of awareness and data processing for top soil quality.
What it does
Predicts topsoils quality and erosion over years using Machine Learning and databases.
How we built it
We started with data pipeline, followed by AI integration and then used bun js to manage all dependencies and worked on the ui.
Challenges we ran into
Prompt engineering, managing parallel runtime for both AI models.
Accomplishments that we're proud of
Clean UI and managing 6 API attachments in 1 project. APIs Used
API Purpose SoilGrids v2.0 (ISRIC) Soil composition — pH, organic carbon, clay %, nitrogen, bulk density NASA POWER Climate data — 30yr rainfall, temperature, humidity, wind averages OpenLandMap NDVI vegetation index from MODIS/Sentinel satellite imagery Nominatim (OSM) Geocoding — converts location text to coordinates Groq — LLaMA 3.3 70B AI risk synthesis — scores degradation risk and generates recommendations GreenPT — green-r-raw Sustainability scoring — carbon sequestration potential and regenerative practices
What we learned
Prompt engineering, parallel runtime for AI, predictive reasoning
What's next for Terravue
Click anywhere on map to get data instead typing, using Wolfram Alpha for universal soil equation.
Log in or sign up for Devpost to join the conversation.