The submission, field by field
November 15, 2026 · 9:00 AM
The Space Apps entry form has three places where the project lives or dies: the project details, the NASA data list and the partner resources list. With the noon deadline, this is what we filed.
Project details
The text we sent, as submitted:
GrazePoint is a web tool that answers the question a cattle rancher asks every morning: which paddock should the herd enter today, and which ones need rest, cutting, or reseeding. We built it around a real case. At Hacienda Patiño in Santa Cruz, Bolivia, that call is made by eye, walking the field, by a technician with years of experience. We interviewed the ranch’s director the day before the hackathon and designed the tool against her answers.
How it works: you draw your farm’s paddocks on a satellite map (or import them), register what you know (forage species, last grazing date, herd size, soil lab results), and the app crosses that with open satellite data. Each paddock gets a re-growth clock: days since last use, weighted by the water and heat it actually had, so a dry cold week counts less than a wet warm one. Satellite vegetation index (NDVI) then confirms the recovery or flags a disagreement worth checking. The output is a ranking: which paddock is at its optimal grazing point, how many days until the next one is ready, which grass has gone past maturity, and a planting window analysis for reseeding. Reports go out as a PDF over WhatsApp, the channel the field crew already uses.
Benefits and intended impact: it turns the trained eye of one technician into a number anyone can check from a phone, measured on their own land rather than a regional average. The immediate beneficiaries are real: Fundación Patiño trains over 300 producers a year and supplies certified seed to 1,342 producers across seven departments, and the ranch itself runs on rotational grazing. Beyond that network, the tool works for any ranch the satellites pass over, for free, because every input is open data.
Tools and stack: a bilingual (English/Spanish) browser app in HTML, CSS and vanilla JavaScript with Leaflet maps, backed by a Node.js REST API at /api/grazepoint with a mirrored offline backend, PDF report generation, and WhatsApp/Telegram delivery. NASA datasets: POWER daily API (rainfall, temperature, top and root-zone soil wetness), SMAP L4 root-zone soil moisture, and MODIS NDVI served through GIBS. Because soil and crop data are not NASA products, we add open partner sources: SoilGrids via WCS, Open-Meteo soil layers, and the ranch’s own lab analyses and biomass measurements.
What makes it creative: the effective re-growth days model, a clock the rancher can read instead of raw indices, and NDVI treated as a referee that can disagree with the clock, with the disagreement shown as information instead of hidden. We also turned a data gap into a feature: Bolivia has no public parcel or rotation data, so the app says so plainly and uses vegetation history as the proxy.
Factors we considered: field connectivity and devices (WhatsApp first, PDF over dashboard), adoption cost (registering a grazing event takes three taps), honest uncertainty (global grids are coarse in the Andes and tropics, so we show ranges and warn when a paddock is smaller than the 250 m pixel), and scientific validity (the model infers maturity, a proxy for lignification, not protein content, and the report says so).
NASA data
Every NASA resource the app uses, linked as the form asks:
- NASA POWER API (Agroclimatology)
- NASA GIBS (Global Imagery Browse Services)
- MODIS/Terra Vegetation Indices NDVI 8-Day (MOD13Q1)
- SMAP L4 Root Zone Soil Moisture (SPL4SMGP)
- MERRA-2 Bias-Corrected Monthly Precipitation
- MODIS Land Surface Temperature Daily (MOD11)
- FIRMS — MODIS Thermal Anomalies / Active Fire
- MODIS Land Cover Type IGBP Annual (MCD12Q1)
Each product in the list is one the code actually reads. The last six arrive through GIBS layers (MODIS_Terra_NDVI_8Day, SMAP_L4_Analyzed_Root_Zone_Soil_Moisture, MERRA2_Precipitation_Bias_Corrected_Monthly, MODIS_Terra_L3_Land_Surface_Temp_Daily_Day, MODIS_Terra_Thermal_Anomalies_All, MODIS_Combined_L3_IGBP_Land_Cover_Type_Annual), so if a judge checks, the names match the source.
Partner and other data
Everything non-NASA, data first and code after:
- SoilGrids — ISRIC World Soil Information (WMS)
- Esri World Imagery basemap
- Esri World Boundaries and Places (reference layer)
- OpenStreetMap
- Fundación Patiño (field interview, forage species catalog, activity reports)
- Telegram Bot API
- WhatsApp Click to Chat (wa.me)
- Leaflet 1.9.4
- Node.js
- Express
- MongoDB
- Mongoose
- PDFKit
- pdf-lib
- PDF.js (Mozilla)
- JSZip
- shpjs (shapefile-js)
- proj4js
- pngjs
- Nodemailer
- node-jsonwebtoken
- cors / dotenv
- Playwright
- Jekyll
- GitHub Pages
Two things worth noting about how the lists were built. The Patiño row is not decoration: the forage catalog and the grazing criteria come from a real field partner, and that is the strongest card for Local Impact. And the list stays honest: Open-Meteo, FAO GAEZ and the rest of the wishlist from the design docs did not make it into production, so they do not appear. Better a shorter list that matches the running app than a padded one.