AVM

Dataset

Description

An automated valuation model (AVM) is an estimation of a property's value based on recent sales of similar properties in the area adjusted for known differences between the properties. ATTOM's AVM achieves a national median absolute error rate of 4.3% when compared to actual sale prices, with 70% of valuations falling within 10% of the arm's-length sale price.

Our enhanced confidence scoring leverages machine learning to provide property-specific uncertainty estimates. Our Forecast Standard Deviations (FSDs) are calibrated to actual market performance — when we indicate 10% FSD, historical data shows 68% of similar properties actually sold within ±10% of our estimate.

ATTOM ID	145695081
Valuation Date	10/15/2024
Estimated Value	$464,234
FSD	12%
Confidence Score	88
Value Range	$408,526 - $519,942

Example

Tables/Views

The following tables can be used to access AVM information. When delivering solutions to customers, some or all of the information found in these tables may be consolidated into one or more views.

AVM

Rows: 94 million
Columns: 10
Family: Valuation

Dimension Tables

There are no dimension tables associated with AVM data.

Understanding the Data

ATTOM ID

Because there is not a universally accepted standard for property identification (like the use of SSN to identify US citizens), properties can be difficult to identify. Alternatives such as APN or street address can change over time and they often lack a consistent way of being formatted. ATTOM addresses this by assigning a single, permanent identifier to each property in the form of a very precise number (integer).

Being numbers (integers), ATTOM IDs can be processed very quickly in relational databases, improving the speed with which properties can be identified and matched.
An ATTOM ID of 99999999 is assigned for properties that we cannot identify/match

Valuation

The following is the key data provided as part of each property AVM:

Estimated Value: The estimated value of the property
Estimated Min/Max Values: The lower and upper ends of the property's value range
Valuation Date: The date on which the valuation was established
Forecast Standard Deviation (FSD): The expected percentage deviation between our estimate and the likely sale price. An FSD of 12% means we expect 68% of similar properties to sell within ±12% of our estimate.
Confidence Score: Simply 100 minus the FSD. A 12% FSD constitutes a confidence score of 88.

Product Articles (P-Articles)

All solution packs containing the AVM data element will include the following p-articles to help you understand and use the data: -

AVM 101