Predicting Car Sale Time with Data Analytics and Machine Learning

Abstract

There is no doubt that marketing is an important step in Product Lifecycle Management (PLM) and obviously decreasing time-to-market is crucial to reduce storage costs and increase profit. This paper aims to improve marketing strategies in the automotive field for car dealers and car selling supply chain. Due to the cost of new cars and the high risk of car value depreciation it becomes necessary for car dealers to know which type of cars can be sold faster than others, this will allow dealers to adapt their marketing strategies and satisfy the need of their customers. We propose to use data analysis and machine learning algorithms to address this problem and create models to help these companies in their decision-making processes. In our experiments, we used sale data from two big dealers of multi-maker cars. The dataset contains the sale history of around 73200 cars over a period of 8 years. We compared the different machine-learning algorithms and got promising results classifying cars into different predicted sale time ranges.