University of Washington Databricks Hackathon

Option 1: Weather-to-Yield Signal DetectionGiven historical yield data (per geography or crop), plus historical weather variables (rainfall, temperature, humidity, wind, soil moisture if available).Identify which weather features most strongly correlate with yield variations.Flag anomalies: counties/fields/seasons where yield is unusually high or lowrelative to weather conditions.Build a model that can explainwhythose anomalies may have occurred.Bonus:Visualize the anomalies on a map (e.g., heatmaps by county).Explain yield deviations during extreme events (2021 drought, flood periods, heat waves, etc.).Option 2: Predict the Future - 60- and 90- day forcast yield simulationBuild a predictive model that estimates expected yield for the current growing season based on near‑term forecast weather.Quantify uncertainty (e.g., “yield could drop 10–18% if rainfall is below X”).Provide early‑warning insights for agronomists or growers.BonusVisualize prediction intervals and uncertainty bands.Provide a simple natural‑language summary for farmers (“If temperatures exceed 95°F for more than 4 days, expected yielddecreases by 12%.”)Recommended DatasetsYields for Corn and Soybean -GitHub LinkHistoric Weather Data (Databricks Marketplace - multiple options)