Every market statistic on this site comes from one of 2 MLS sources, and each number has exactly one. County and Metro Atlanta figures come from GAMLS InfoSparks. City, zip code, and subdivision figures come from our own closed-sales archive, compiled directly from FMLS (First Multiple Listing Service). We do not source statistics from consumer portals or third-party aggregators. What you read here is the same data we use to price listings and write offers.
Which source does each number come from?
- County and Metro Atlanta pages: median sale price, homes sold, homes for sale, months of supply, days on market, and sale-to-list price come from GAMLS InfoSparks, single-family homes, compiled monthly by the Terence Richardson Group. Price, days on market, supply, and sale-to-list are GAMLS 12-month rolling figures. Homes sold and homes for sale are counts for the month shown. Metro Atlanta means the 12 core counties GAMLS reports together.
- Cities, zip codes, subdivisions, and failed listings: FMLS closed and terminated records, because GAMLS does not publish them at those levels. County pages also carry FMLS price per square foot and seasonality, labeled as FMLS.
- A county page never shows a GAMLS figure and an FMLS figure for the same measure. A build check fails the site if any page shows two different values for one measure of one place.
How is the market temperature scored?
The TRG market temperature is a 0 to 100 score for how warm a county's single-family market is compared with its own history. It compares the current month with the same calendar month in every year GAMLS has on record, back to 2014, so the season never moves the score.
- Take 3 GAMLS measures for that month in every year: months of supply, median days on market, and sale-to-list price. Lower supply, fewer days, and a higher sale-to-list price count as warmer.
- Rank each year against the others on each measure, and average the 3 ranks into one composite per year.
- The score is where this year's composite falls among all the years: 100 is the warmest of those months on record, 0 the coolest. Ties count half.
Bands: 0 to 19 Cold, 20 to 39 Cool, 40 to 59 Typical, 60 to 79 Warm, 80 to 100 Hot. A place needs at least 5 years of that month on record before it gets a score. The temperature describes the market, not any one home, and it says nothing about where prices go next.
Where does the data come from?
From more than 1 million FMLS closed residential sales across the Atlanta core counties, with continuous coverage since January 2010. Henry County carries history back to 2008. The archive refreshes on a rolling basis. Each refresh re-pulls a trailing 60-day window, so late-reported and restated closings are absorbed. Each monthly site update then recomputes every geography page from scratch. Current coverage runs through 2026-10-07, and pages state their data vintage as September 2026.
What do city, zip, and subdivision pages show?
6 metrics, all computed from FMLS closed sales:
- Median close price: the median closed sale price over the trailing 12 full months (2025-10-01 to 2026-09-30 on the current build).
- Closings: the count of closed sales in that same window.
- Median days on market: the median of FMLS total days on market for those closings.
- Median price per square foot: computed only on closings that report usable square footage. Records reporting zero or implausible footage are excluded, not guessed.
- Average close-to-list ratio: the mean of sale price divided by original list price (SP/OLP). Data-entry outliers outside 50% to 150% are excluded.
- 5-year trend: annual median close prices for the last 5 full calendar years. Years with fewer than 5 closings in a geography are omitted, not plotted.
How are geographies defined?
- The site has 2 geographic surfaces. Market Data covers the full 12-county region at county, city, and subdivision level. Communities covers only the areas that have a relocation guide. Both read the same dataset.
- Fulton County is published as 1 county with a 3-region breakdown. North Fulton is north of the river. Fulton ITP is inside the I-285 Perimeter: Atlanta proper plus East Point and Hapeville. South Fulton is outside the Perimeter to the southwest. Atlanta-addressed records in Sandy Springs zips (30328, 30350) count as Sandy Springs. Records in unincorporated South Fulton zips (30349 and neighbors) count toward the South Fulton region only. Zip 30331 straddles the Perimeter and is counted ITP. Where a zip crosses a line, we assign the whole zip rather than guess at parcel level.
- Some pages cover named submarkets that are not municipalities. East Cobb means Cobb zips 30062, 30066, 30067, and 30068. Buckhead means City of Atlanta zips 30305, 30324, 30326, 30327, and 30342. Vinings means Cobb zip 30339. South Forsyth means zips 30004, 30005, 30024, and 30041. The Lake Lanier corridor means Forsyth zip 30506 plus Gainesville-addressed Forsyth records. Each page states its definition.
- “Atlanta” means the City of Atlanta: records in Fulton and DeKalb counties. The Sandy Springs and unincorporated South Fulton zips above are excluded. Cobb-side records with Atlanta mailing addresses in 30339 count as Vinings.
- Months-of-supply, days-on-market, sale-to-list, and price figures at metro and county level come from GAMLS InfoSparks, compiled monthly by the Terence Richardson Group. They carry that source label wherever they appear.
- Each city page lives under its dominant county. It includes all of that city's closings, including slivers that cross county lines.
- Subdivision pages exist for every community with 10 or more closed sales in the last 12 months. A community below that is reported on its city page. Name variants (phase and unit suffixes, inverted articles) are normalized into one canonical community before counting.
How is the failed-listing rate computed?
A failed listing left the market Withdrawn or Expired without closing. The failure rate for a period is failed divided by failed plus closed, among listings terminating in that period. A home that was relisted and later sold appears in both series: the rate measures listing attempts, not homes. The series comes from FMLS MASTERSTATS withdrawn and expired records back to 2010, compiled by the Terence Richardson Group. County-years without reliable failed coverage (Henry before 2026) show no rate rather than a false low one, and no rate is published on fewer than 20 terminated listings.
When is a sample too small to publish?
Medians display from 5 closed sales and carry a visible small-sample caution below 12. Question-and-answer text and extractable claims quote medians only at 12 or more sales; below that they cite the last full year with enough volume. Comparison claims (close-to-list ratio, failure rates, trend deltas) require 20. Sample size displays on every stats block. Below every threshold we roll the reader up to the parent geography instead of guessing.
What gets excluded?
- Sales below $10,000 and above $60 million. Those are data-entry errors.
- Records without a parseable close date.
- Off-MLS transactions. FMLS coverage is deep but not total. New-construction and investor bulk deals that never list are not represented, and we say so rather than estimate them.
- The southern-arc counties (Henry, Clayton, Fayette, Rockdale, and parts of South Fulton) trade meaningfully on GAMLS as well as FMLS. FMLS-based figures there capture a large share of the market, not all of it. Henry is the clearest case, and its pages say so. Trends in these counties are directionally reliable. Totals understate whole-county volume.
Where does national context come from?
From 4 primary sources, linked at the point of use: Freddie Mac for mortgage rates, the U.S. Census Bureau for population counts, and NAR and FRED for national housing figures. Local statistics never come from third-party aggregators.
Questions about the data, or want a cut we haven't published? Ask us directly. Attribution for citations: “GAMLS InfoSparks, compiled by the Terence Richardson Group” for county and metro figures, and “FMLS closed sales, compiled by the Terence Richardson Group” for everything else.
