A typical home seller in the U.S. in 2017 was 56 years old. They had lived in their house for the last ten years and had a median income of 107 thousand dollars. While the median age of a first time home buyer was 33, and 47 for a repeat buyer, with respectively 80 thousand dollars and 106 thousand dollars of income. The standard home bought was 1,900 square feet in area, was constructed in 1993, and had three bedrooms and two bathrooms. And on average, 88% took a loan to finance their purchase. Needless to say that, for such a substantial debt contracted, the actual price negotiated has to be as close as possible to the fair market value.
We have used Quaartz with a dataset describing 409 transactions. It is a regression use case. The following information is included for each transaction:
- Price per square meter (goal of the prediction),
- House street number,
- Last transaction date,
- House age,
- Distance to nearest metro station,
- Number of convenience stores,
- Latitude,
- Longitude.
The goal is to see whether Quaartz can estimate the home selling price accurately. The price estimation has been mostly dependent on the realtor’s evaluations. However, it is precious for a home seller and a realtor to get a quick, independent, objective estimate in a few clicks.