|
|
Predicting the Selling Price of a Used Car
marketing icon

Predicting the Selling Price of a Used Car

Sales of used light vehicles in the United States grew to about 40.8 million units in 2019. The same year, the average selling price of used cars came to nearly 21,000 U.S. dollars. TADA can help sellers anticipate a used vehicle's selling price before it is put on the market.

Industry

Car

Project Duration and Effort

One week

Type of Prediction

Regression

Customer Benefits

  1. Find in minutes the fair market value of a used car with a reasonable accuracy.
  2. Understand which parameters influence the price.
  3. Anticipate the yearly price drop for a specific car.

Problem to solve

In the United States, the average selling price for a new light vehicle was approximately 36,800 U.S. dollars in 2019. New automobiles and light trucks were on average almost 16,000 U.S. dollars more expensive than used light cars: the average selling price for used vehicles amounted to about 20,840 U.S. dollars in 2019. In light of a growing price gap between new and used cars, an increasing number of U.S. households will likely choose a used car over a new one.

The issue is whether or not it is possible to estimate the selling price of a used car. We have used a dataset comprising 120 records describing 120 used vehicles. The descriptions contained 24 different fields of information, which can be summarized as follows:

  • Characteristics related to the vehicle’s market segment, i.e., premium or mass market, normalized yearly loss,
  • Physical and mechanical characteristics of the vehicle, i.e., width, weight, engine type, and several other fields for each group of characteristics.


Current approaches for used car price computation are basic and calculate a rough average, which does not account for a particular used car’s specificities. Instead of these outdated solutions to get a listing price for a used car, TADA can compute an exact price estimate.

Objectives

  • Based on few records, find the fair selling price of a used car.
  • Understand the criteria that impact the price yearly price loss for a specific car.

 

It raises the following question:

Is it possible to predict the selling price of a specific used car?

Solution

In this regression use case, TADA can predict a used car’s price with an R2 of 88%, a MAPE of 11%, and an RMSE of 1769.

Buying a used car is the number one choice in most western countries. There are about 280 million vehicles in operation in the United States, an increase of about 1.6 percent. The rising demand for used cars means that used vehicle inventories are declining. As a result, used-vehicle prices are expected to increase.

Financial reports have exposed how the coronavirus pandemic outburst has triggered a change in vehicle-buying behavior in the United States. With various consumer goods and services presently purchased online due to Covid-19, the auto industry has also begun to integrate its services online to reach consumers digitally. This trend makes it all the more essential to have simple access to used car price estimations.

TADA has selected the following five main criteria out of the twenty-four available in the dataset:

  • Vehicle curb weight, with an impact of 27% in TADA’s decision,
  • Vehicle width, with an effect of 27% in TADA’s decision,
  • Vehicle body style, with an effect of 17% in TADA’s decision,
  • Market segment, with an effect of 17% in TADA’s decision,Normalized yearly loss, an effect of 10% in TADA’s decision.
It is interesting to note that the most influential criteria are tightly related to the car’s physical and mechanical characteristics, i.e., vehicle body style, width, and not to its market price. The normalized yearly loss has a relatively small influence on the price estimate.

Live Predict

Customer Benefits

In one week, car sellers managed to build a used car price estimator with an excellent R2:

  • Get a fair used car price estimate with an R2 of 88%.
  • Understand that car characteristics make its price and found which features did.