In this project, we trained a series of relevant ML models including KNN and tree-based methods on thousands of observations obtained from web scraping a vehicle reselling website. The performance of the models were evaluated using a test sample, and when possible, most important features for determining the vehicle resale price were determined. Finally, we explored the implications of our findings and discussed them using economic theory.
You can find a presentation of our contribution, our analysis code, web scraping code, produced dataset and final paper attached.