Nice data, but nearly impossible to understand. It would be nice to put at least a couple of big cities (Palo Alto, Fremont, etc) on the map, so that you could understand where's what.
Under Debian, US Census TIGER files are available as dictionary files. You can use this to identify specific areas as you wish.
My analysis showed that the most expensive top several ZIP codes were SF Financial district, Stanford, SOMA, Marina, and Pacific Heights neighborhoods, as well as Atherton. I suspect the data may skew high, though it would help to have a distribution within each ZIP code as well (SD, median, quartiles, 5th/95th percentiles).