Your question is about the distribution of the data between the current batch of sales and the prior batches of sales that contributed to their respective medians.
Price/SqFt is not that useful because the “price” part of it includes the value of the site as well as the improvements, which can vary tremendously. For example, a 2,500 SqFt home on a 6,000 SqFt lot that sold at $600,000 would result in a $/SqFt indicator of $240; You’d get the same indicator off a 4,500 SqFt home on a 15,000 SqFt lot but that sold at $1,080,000. The $/SqFt indicators are the same but the effects of those sale prices on the medians would be much different.
The way I’ve been analyzing data distribution among the various zips is to separate them using age ranges as the primary and size ranges as the secondary. That works pretty well, but it’s labor intensive.
