Transparency
Methodology
OpenRoof derives several composite metrics from public data sources. This page documents exactly how each metric is constructed, what inputs drive it, and where we suppress results rather than display low-confidence numbers.
Valuation
Peer Fair Value
We find comparable ZIP codes within the same metro area that share a similar median price tier, then run an Ordinary Least Squares (OLS) regression of median home value on median household income across that peer group. The predicted value at any given income level is what we call peer fair value.
The gap between actual and predicted value is expressed as ov_vs_peers_pct, showing how overvalued or undervalued a ZIP is relative to its structural peers — not relative to a national average.
Conceptual formula
fair_value = β₀ + β₁ × income + ε
Per-metro OLS · peer group only · computed quarterly
Confidence threshold
If the R² of the regression is below 0.3 — which occurs when the peer group has fewer than 8 members or when income explains little of the price variance — we suppress the result entirely. We prefer showing no number over showing a misleading one.
Important caveat
This is not an automated valuation model (AVM). OpenRoof does not estimate individual home values. This metric reflects how a ZIP compares to its peer group on structural fundamentals — income relative to price — and should not be used as a substitute for a professional appraisal.
Attribution: OpenRoof OLS model, computed quarterly. Peer groups drawn from Zillow ZHVI + U.S. Census ACS.
Distress
Distress Opportunity Score
The Distress Opportunity Score is a weighted composite of five county-level inputs that collectively indicate the concentration of financially stressed properties in a market. A higher score means more distress — and therefore more potential investor opportunity.
| Input | Weight | Source |
|---|---|---|
| Foreclosure Rate | 30% | County Records |
| REO Rate | 20% | County Records |
| Distressed Listing Share | 20% | County Records |
| Foreclosure Starts YoY | 15% | County Records |
| Tax Delinquent Rate | 15% | County Records |
Handling missing inputs
When fewer than all five inputs are available, the weights of present inputs are rescaled so they sum to 1.0. When fewer than 2 inputs are available, the score returns null and is not displayed rather than showing a misleadingly low reading based on a single signal.
Current coverage note: OpenRoof uses public county records for distress data. Signal availability varies by county: sheriff-sale sources can provide foreclosure activity, while tax sources provide delinquency rates. Scores are only displayed where at least two inputs are present. See coverage →
Important caveat
This is not a buy recommendation. The score reflects current stress levels only. Coverage varies by county — see the coverage map for which counties have live data. Individual due diligence is required before acting on any distress signal.
Peer Matching
Similarity Score
The Similarity Score powers the “Similar Markets” feature, which surfaces the five closest-scoring ZIP codes in the same metro. It uses a four-dimensional peer scoring approach based on Euclidean distance.
Dimensions
- Median home price
- Year-over-year % change
- Days on market
- Rent yield
Computation
- Each dimension is normalized to the range
[0, 1]across the full metro dataset. - Euclidean distance between the target ZIP and every candidate ZIP is computed in normalized 4-space.
- Similarity is computed as
1 − normalized_distance, clamped to[0, 1]. - The five ZIPs with the highest similarity score within the same metro are returned.
Data Sources
Attribution & Provenance
Every MetricCard in OpenRoof shows a provenance indicator on hover that surfaces the data source, last-updated date, and refresh cadence. The table below lists all primary sources used across the platform.
| Source | Data | Cadence |
|---|---|---|
| Zillow ZHVI | Home values (ZHVI) | Monthly (mid-month) |
| Zillow ZORI | Average rent (ZORI) | Monthly (mid-month) |
| Realtor.com | Inventory, days on market, price cuts, new listings | Monthly (1st week) |
| U.S. Census ACS | Median income, population, demographics | Annual (Q4 release) |
| FRED | Mortgage rates | Weekly |
| County Records | Tax delinquency rates | Varies by county |
| County Records | Sheriff sales, foreclosure activity, REO indicators, lis pendens where published | Varies by county |
| OpenRoof OLS | Peer fair value estimates | Quarterly |
Freshness labels
Every chart and map tooltip displays a “Data as of [Month Year]” label so you always know the vintage of what you are seeing. Stale data is never displayed without this label.
NULL handling
A NULL value in any metric means no data is available for that location at that point in time. NULL is never displayed as zero — doing so would imply a meaningful reading where none exists.