Title:

Gas and Oil Usage Patterns/Predictions

Palmer Gas and Oil is a gas and oil distributor across the New England area. With customer’s previous order data, they were looking for an efficient and accurate way to help customers determine how much they should order. In collaboration with SilverTech, Inc., the company behind their website, our group was tasked to create a machine learning model that predicts the amount of oil a customer will need for the next two months and a visualization of those predictions. The predictions are based on previous order history, obtained from the customer database, and current weather trends in the customer’s area, provided by a national climate database. Customers can view their predicted oil usage via a dashboard on their customer portal. This graph allows customers to view their previous purchases and predicted purchases by month and year. The graph has several overlay options that help to show how temperature data affects a customer’s gas usage. The model will continue to update with each passing month, with new order history and weather data to ensure that customers predictions stay accurate and reliable. We tested our model for accuracy by excluding the final year of historical data from its training period and then comparing that data to our model’s predictions for that year.

Name

Comment*

Comments are viewable only by submitter

Captcha*