Put On Your Data Mining Helmets

As regular Piggs know, I routinely chart out either the size-adjusted median home sale price (which is timely, but not too accurate) or the Case-Shiller home price index (which is more accurate, but not very timely). In this post I’d like to highlight a couple of folks who have been making an effort to better measure home prices without such a lag.

First, our very own forum participant "esmith" has collected the data and written the code to roll his own version of the Case-Shiller methodology, except that he uses the most recent month’s sale prices and is able to break out some San Diego submarkets. He recently posted his January data — have a look at the charts and accompanying discussion, and if you need an introduction to his technique you can find it here.

Second, Bruce Henderson at "And Still I Persist" has written an application that models prices by mining the MLS. Bruce uses MLS asking prices instead of sale prices, which may introduce some noise but also allows the price data to be captured pretty much in realtime (because a closed sale represents a price that was agreed upon 1-2 months earlier). The aforelinked article is chock full of charts showing his results for various SD submarkets.

As if often discussed here, no single indicator is perfect, but being able to look at a whole bunch of them at least allows us to narrow down what’s actually going on. So thanks, Bruce and esmith, and keep that data coming!