Side Project - Predictive Real Estate Timing Website

I've decided to frantically catch up with the runaway train called Web 2.0. One way of getting up to speed is by making a project out of Web 2.0 technologies.

Although there's been a *tremendous* amount of activity in my life lately, along with an innumerable number of side projects, the very first project that I want to personally tackle, with my statistics guru-buddy, is simple: a proven, near-real-time predictive engine for real estate valuation. Basically, this will be a tool that'll help the average Joe have an idea about real estate market timing.

The core engine, which includes automated mass data feeds, modeling, and algorithms, was completed a year ago, with some incremental fine-tuning every so often. Privately, my stats buddy and I have been doing very informal validation testing of the engine, by correlating the uptrend, peak, and starting downtrend of real estate movement, and corroborated with anecdotal observations-- my own for one, by manual monitoring and correlation of median prices in Southern CA MSA's to the sale of my own condo, a few heated arguments between him and me about the viability of the Texas income property market (suffice to say, he won that argument), and news articles and blogged experiences we've been witnessing wide-eyed.

Features will include: correlation with the Case-Shiller Home Pricing Index, backtesting, forecasting and trending by ZIP code, etc.

This is significantly different from what Zillow.com attempts to portray. Zillow.com more or less is a crude quickie "online appraiser" for individual properties. Although their algorithms are also proprietary, it's safe to assume that Zillow.com may be using the 3 standard pricing approaches of real property: comparative (at what prices are other recently sold properties?), income (at what price is considered fair, given amount of rent commanded by property?), and functional replacement (to rebuild property from scratch + depreciation, how much is property worth?) estimation approaches.

Zillow's restrictions seem to include the following: no future, predictive confidence; estimates seem to be based on final sales figures (*not* repeated downwardly revised *listing* prices); and not factoring in alternative ways of property listing that sellers are employing during these tough times; no accounting for economic conditions, which may be significant factors on real estate values.

So the engine's there. What's missing, then? The web front-end! And obviously considerations for payment processing and security will be high priorities as well.

Stay tuned.