Core answer
Historical rain probabilityFor one selected location and calendar date.
Built for NASA Space Apps Challenge
Willie Rainy
Rainfall climatology from historical data.
A five-person hackathon project estimating how often measurable rain has occurred at a selected location on a selected calendar date.
Mission question How often has it rained here on this date?
01 / Mission
Mission
Useful context for rain-sensitive plans beyond the reliable forecast window.
Travel, farming, outdoor events, and seasonal decisions still need context when long-range forecasts cannot provide certainty.
Willie Rainy reframes that gap as a climate-history question, showing what the archive says without presenting climatology as prediction.
02 / Method
Method
A transparent path from a planning question to an uncertainty-aware result.
- 01Location + Date
- 02Historical Rainfall Archive
- 03Calendar-Day Grouping
- 040.1 mm Rain Threshold
- 05Probability + Wilson Interval
- 06Climatology Result
03 / Prototype
Prototype
Select a place and date, then read the historical signal with its uncertainty intact.
- 01Location search
- 02Browser geolocation
- 03Manual coordinates
- 04Date selection
- 05Result view
Uncertainty
95% Wilson intervalA visible range that avoids false precision.
Rainfall
Mean + medianCalculated from rainy observations only.
Context
Sample + sourceHistorical sample size and methodology metadata.
Loading, validation, network, and provider error states keep the experience legible when a request cannot be completed.
04 / Build
Build
A typed interface, asynchronous API, and historical rainfall service.
Interface
The React and TypeScript frontend handles location input, calendar-date selection, result presentation, source metadata, and user-facing request states.
API
FastAPI and Pydantic validate latitude, longitude, date, and range inputs before passing each request to the climatology service.
Data method
An asynchronous aiohttp client retrieves daily rain_sum values from Open-Meteo, groups them by calendar day, and calculates the statistical result.
User browser
React frontend
FastAPI backend
Climatology service
Open-Meteo archive
Result UI
Open-Meteo historical weather and geocoding APIs supply the archive and location data. An in-memory cache reduces repeat requests.
Tech stack
- React 19
- TypeScript
- Vite
- Tailwind CSS
- Axios
- FastAPI
- Pydantic
- aiohttp
- NumPy
- Open-Meteo
- Docker Compose
05 / Role
Role
Full-stack implementation and product direction within a five-person team.
- Built parts of the React and TypeScript query interface.
- Helped implement single-day and date-range FastAPI endpoints.
- Integrated Open-Meteo historical weather and geocoding data.
- Supported rain probability, rainfall summary, and Wilson interval calculations.
- Improved validation and error handling across frontend and backend.
- Contributed to provider structure and the Docker Compose setup.
06 / Outcome
Outcome
Historical context without overstating prediction.
The prototype turns raw rainfall observations into a practical planning interface and demonstrates collaborative delivery, climate-data handling, full-stack API design, and uncertainty-aware product thinking.