Skip to content

Aryan-work/Databricks-Hackathon

Repository files navigation

🌾 Databricks Hackathon Submission

Aryan Sharma • Daniel Zhang • Varun Atraya

🎥 Demo Video:
https://youtu.be/Ezpgv-hSWfg


🚜 Business Summary

Raw yield numbers can be misleading.

A strong year may simply reflect favorable weather.
A weak year may be driven by drought or heat stress beyond anyone’s control.

Our system establishes a weather-adjusted yield benchmark.
We estimate the yield expected under specific weather conditions and measure the deviation from that baseline.

This weather-normalized signal separates:

  • 🌦️ Environmental impact
  • 🌱 True land or management performance

The result is a clearer view of structural productivity and resilience.


💼 For Farmland Investors

Historical averages do not reveal whether land is intrinsically productive — or simply benefited from good weather.

By identifying consistent overperformance or underperformance relative to weather conditions, our platform enables investors to:

  • ✅ Detect resilient, high-quality land
  • 💰 Avoid overpaying for weather-inflated yield history
  • 📉 Assess long-term climate risk exposure
  • 📊 Compare assets on a normalized performance basis

This supports smarter acquisition, valuation, and portfolio allocation decisions.


🌽 For Farmers

Year-to-year yield comparisons are often distorted by weather variability.

Weather-normalized benchmarking allows farmers to:

  • 📈 Evaluate performance fairly across seasons
  • 🔍 Distinguish weather-driven losses from operational gaps
  • 🌾 Demonstrate resilience and management strength
  • 🧠 Identify structural improvement opportunities

This transforms raw yield history into a performance story backed by data.


📊 Core Metric

Yield Residual = Actual Yield – Weather-Predicted Yield

  • 🟢 Positive residual → Outperformance
  • 🔴 Negative residual → Underperformance
  • Near zero → Yield aligned with weather expectations

This simple but powerful metric converts historical yield and weather data into actionable performance intelligence.


🔬 Modeling Framework

We train crop-specific models using historical yield and weather data from 2010–2024.

For each county and year, the model outputs:

  • Predicted Yield (based solely on weather conditions)
  • Actual Yield
  • Residual (performance signal)

Model Performance

  • RMSE: 18.55 bu/acre
  • R²: 0.868

This demonstrates that weather explains the majority of yield variability while preserving a meaningful structural signal.


🌎 Why It Matters

By isolating weather effects, we unlock:

  • Climate-adjusted land valuation
  • Resilience identification
  • Drought and extreme event analysis
  • More accurate long-term productivity assessment

In a world of increasing climate volatility, understanding true performance — independent of weather — is a strategic advantage.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages