Automated Benchmarking for the Purchase of Second-hand Cars and Motorcycles in Colombia through Web Scraping and Prescriptive Analytics

Abstract

In the current context of the automotive market in Colombia there is an exorbitant rise in prices of second-hand vehicles and in the absence of structured tools to access quickly and reliably to the market information, it is necessary to create a centralized dataset that reflects the inventory offered on the web, in order to facilitate the comparison, collection and decision making in the process of buying and selling second-hand vehicles. Under this premise, this applied research article proposes the creation of a minimum viable product that allows to extract, prepare, process, transform and compare the information via Web Scraping and propose decision alternatives using prescriptive analytics based on decision trees and benchmarking for users of the second-hand car and motorcycle market in Colombia. The results of the developed tool are satisfactory since they offered the user two options according to the sought characteristics of the motor vehicle, one of minimum price and the other of minimum mileage-lowest price. Finally, to go further, it is recommended to explore more complete frameworks for a more personalized experience.