![IconGems-Print](images/IconGems-Print.png)
121 – SPEED: Data Expo
An Analysis on the Accuracy of Weather Forecasts
Benjamin Schweitzer
Miami University
Robert Garret
Miami University
Thomas J. Fisher
Miami University
One of the most popular usages of predictive modeling is in weather forecasting. We aimed to utilize machine learning techniques to extend provided forecast to sites across the continental United States. The forecasts and observed weather for 113 sites across America (2014 - 2017) were supplied to us by the competition officials. Supplementary data on observed weather was acquired from the National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center. We then created visual displays to analyze the prediction accuracy of the extended forecasts. Our results allow for an in-depth exploration into the accuracy of weather forecast across the nation.