Inspiration
Since we care strongly about the environment and the climate change crisis, we wanted a more efficient way to predict the weather based on previous years and ongoing trends.
What it does
It predicts the temperature of the inputted date (from the user. It can predict the temperature (min, max), and precipitation (inches) for future dates.
How we built it
Our model uses a RandomForestRegressor with automatically created decision trees to parse the data, then we used gridsearch_cv to optimize the RandomForestRegressor model in order to minimize error. We are returning the result of the best model using best_estimator_.
Challenges we ran into
Initially, we used linear regression, but its accuracy score was very low, because our data had a convex shape. We tried RandomForestRegressor, which gave us a very low error for all of our tests. Although running our model took a long time (since our data was almost 7000 lines), our predictions were successful.
Accomplishments that we're proud of
We were able to create a successful ML model that could predict weather for future dates, and our predictions made sense in the context of previous dates.
What we learned
We learned how to use the RandomForestRegressor as our ML model to make predictions based on given data, and gridsearch_cv to optimize our model.
What's next for San Jose Weather Predictor
We could also add data and make predictions for other weather patterns, such as wind speed and direction, humidity, air quality, and visibility.
Log in or sign up for Devpost to join the conversation.