Business Case for Predictive Analytics
Task : Create a machine learning model that correlates used car process from a large dataset of vehicle attributes.
Action : Using the provided data, we were able to create an extreme gradient boosting model that allows us to interpret the data and find patterns to create a prediction on used car pricing. The data set consisted of attributes such as drivetrain, fuel economy, make, model, vehicle age, mileage etc.. Using these attributes the model was created.
Result : The model was successfully created to show the correct pricing with an uncertainty of +- 5000 dollars. It also showed us the most influential factors that come into predicting used car pricing. It turns out that after the make and model of the vehicle, the Drivetrain was the most influential factor for non luxury brand vehicles while engine size to be the most influential for luxury brand vehicles.