Being in the market for a new car is overwhelming! A car is a big purchase, and if you’re like me this is a decision that you only make once every decade or two. There are thousands of cars available so how can you know you made the right choice?
Well, this time I decided to do some data mining on Cars.com to help me figure out what is going on with new and used cars in the area. A little html parsing magic later, I was able to survey all of the new and used cars available in my target class: midsize sedans.
Looking out from new cars to 20 yr-old cars in the class I was expecting some sort of exponential decay in price versus the age of the car. Instead I think the data shows three distinct regions, where depreciation values are stated as a function of the original selling price:
- fast initial depreciation at about 18%/yr over the first two years
- depreciation at about 6%/yr over the next ten years
- stabilization at selling price of $5000
For any given make/model/year/trim category, there is still a pretty wide spread in dealer list prices. My strategy was to watch this data for several weeks and look for outliers – cars that were selling at the tail of the distribution but with clean histories and low mileage. A little data mining is a small price to pay for confidence in car-buying decisions!
Pingback: Cars.com Data Miner | Gregory Kimball