I didn’t run any serious statistical tests on the data set. This is because there are a few fundamental problems, the largest being that we’re not comparing apples to apples in terms of what’s being sold. That is, we don’t know if the same types of bikes are being sold for more in Seattle than in Phoenix, or if there are different types of bikes being sold in the two markets. The ads also change so frequently that replicating these results might be difficult; and 50 data points might be too small a sample.
Still, whether it’s over/underpricing or just selective selling, what struck me about this informal little analysis was that not one city fell out of line in the inverse order. Where cars were selling for the most, bikes were selling for the least; where cars were selling for the least, bikes were selling for the most; and so on, inversely, in between.
