Predicting Yield based on Weather

Development Process

Our group was formed very late in the process, as our group members quit the hackathon halfway through. So, two groups were merged. We tried and struggled to find a weather dataset. That was the most time-consuming part of the process, as clean, good data is hard to come by. Eventually, we found the NOAA website and requested the data for the top 5 states for corn/soybean production. Aggregating the data and cleaning it up also took up a good chunk of time. Then, with the limited time left, we predicted a yield based on the weather conditions, and recorded the difference between actual and predicted, and indicated if it was an anomaly.

Databricks

We had trouble with Databricks because it would often not load and was kind of slow, but the marketplace was very convenient! I can see it being useful in the future, since now we are more familiar with it.

Built With

• matplotlib
• pandas
• python
• scikit-learn