WA created house level indexes and forecasts by treating each house as a unique asset with an individually calculated index. WA produces each index with a database of repeat sales of thousands of houses within the local market. However, unlike other index providers, instead of pretending that each house in the market is identical and producing one index for the entire market, WA overlays this data with characteristics such as the home's age, size and number of bedrooms from public records and the MLS.
WA then focuses the repeat sales data on each house by weighting the sales of the other houses based on their similarity to the subject house. The resulting indexes distinguish value changes even for houses that are next door neighbors. For example, a smaller newer house might be in high demand while the house next door is a larger older house in lower demand. The WA index for the first house will rise while its neighbor will fall.
WA creates visualizations of the changing value of every house in the market to understand pricing patterns at a deeper level and produce uniquely accurate home price forecasts. These visualizations show that market shifts can begin with a small cluster of houses which grows in number and can then expand like a weather front moving across a market. Houses that have not changed value direction can be predicted to shift by observing the changes in the other houses in their proximity. Using this method, WA accurately predicts 90% of the cases when houses that have been appreciating for multiple years cease appreciating in the following year.