Johnston County, North Carolina — the farmland southeast of Raleigh that spent the last decade turning into subdivisions — carries the fastest-growing road traffic in the state. Vehicle counts on its state-maintained corridors compound at +1.99% a year. Union County, on the Charlotte side, runs +1.71%. Henderson and Franklin both sit at +1.67%.
At the other end of the same state, Washington County loses 1.57% of its traffic every year. Tyrrell loses 1.29%. Hyde, 1.11%.
Those are not survey estimates or model outputs. They are what happens when you take the vehicle counts state transportation departments physically collect at fixed points on the road network, line up each station's own multi-year history, and measure the slope.
Corridor momentum, county by county
This is the first run of PropRaven Corridor Momentum (v0), built from 1,041,762 traffic count stations we collect across all 50 states. Measuring growth requires a panel — the same station counted in more than one year — and 147,600 stations clear that bar and the quality guards described below. North Carolina publishes its counts with county attribution, which makes it the first state we can rank at the county line: 73 counties, every one of them measured from its own stations.
Fastest-growing corridors
| County (NC) | Traffic growth, annualized |
|---|---|
| Johnston | +1.99% |
| Union | +1.71% |
| Henderson | +1.67% |
| Franklin | +1.67% |
| Randolph | +1.65% |
| Harnett | +1.53% |
| Hoke | +1.49% |
| Watauga | +1.34% |
| Wake | +1.34% |
| Gaston | +1.17% |
Declining corridors
| County (NC) | Traffic growth, annualized |
|---|---|
| Washington | −1.57% |
| Tyrrell | −1.29% |
| Rockingham | −1.15% |
| Hyde | −1.11% |
Read the top table as a map and it draws itself: Johnston, Franklin, Harnett and Wake are the Raleigh commuter ring; Union, Gaston and Randolph are the Charlotte and Piedmont Triad ring; Henderson is metro Asheville; Watauga is Boone. The counties losing traffic are the rural northeast and the coastal sounds. Nobody told the model where the growth rings were. It found them in vehicle counts.
At the state grain, the same measurement over the twelve other states with a multi-year count panel:
| State | Traffic growth, annualized |
|---|---|
| Utah | +1.96% |
| Arizona | +1.82% |
| Mississippi | +1.28% |
| Maryland | −0.83% |
| West Virginia | −1.20% |
Sun Belt corridors up, Appalachian corridors down — which is the direction anyone watching those regions would expect, and exactly the kind of result a new signal has to get right before its more surprising outputs earn any attention.
Why count the cars
Traffic counts have a property most demand indicators lack: they are a physical census, not a sample or an estimate. A state transportation department puts a counter on a road and records how many vehicles cross it. There is no respondent to misremember, no listing to go stale, no seller to talk up the neighborhood. The count is the count.
They are also early. A corridor gains traffic when people start living, working, and driving along it — which happens as houses fill, not when a deed is later recorded, an assessment is later revised, or a survey is later fielded. The road registers the change first, and the paperwork catches up.
The catch is that a single year's count tells you how busy a road is, not whether the area around it is growing. A four-lane arterial carrying 58,000 vehicles a day is busy; the useful question is whether it carried 46,000 ten years ago or 61,000. Only the history of a station is a growth signal, and only some of these count programs publish one. Where they do, every station becomes its own small time series.
The cross-check
A new signal is worth exactly as much as its agreement with something measured independently, so we ran two.
North Carolina's transportation department publishes its own five-year growth rate on each station. Our figure is computed differently — a ten-year regression slope on log counts, not a point-to-point rate — and from the raw count histories rather than from their published number. Across 23,232 North Carolina stations, the two rank the state's corridors at a Spearman correlation of +0.683. That is strong agreement across a genuine methodological gap, which is the useful kind: it means the growth math is sound rather than merely reproducible.
The second check is more interesting, because it comes from a completely different corpus. PropRaven's recorded-deed archive — county recorder documents, collected from the counties themselves — yields a transaction velocity ranking for the same North Carolina counties. Its leaders include Gaston and Union. So do the traffic corridors'. Two measurements with nothing in common — one a vehicle counter on a state highway, one a stack of deeds stamped at a courthouse — independently name the same counties. That convergence is the strongest evidence in this piece that both are measuring real demand.
Method
Each station's counts are converted to a single growth rate by regressing the natural log of the count on the calendar year over a trailing ten-year window. The slope is an annualized log growth rate, which reads as a continuous compound rate; using a regression rather than first-to-last arithmetic means a station whose program skipped a year — several count programs run on a biennial cycle — is still measured correctly from the points it does have.
Station rates roll up to a county or state by a weighted mean, weighted by the log of the station's most recent count. The weighting matters: an interstate should count for more than a residential collector, but not so much more that one freeway determines a county. We also carry the unweighted median and the share of stations growing alongside every published figure, so a ranking driven by a handful of outliers is visible rather than hidden. Rankings are standardized within grain type — counties against counties, states against states — and expressed as a 0-to-100 percentile.
The guards are deliberately blunt. A count year must fall between 1985 and the present. A station needs at least two dated observations spanning at least two years, and its most recent count must be at least 200 vehicles per day; below that, a single new driveway is a double-digit percentage. Growth rates beyond ±50% a year are dropped outright as recording artifacts, and what survives is winsorized at the 1st and 99th percentiles within its state.
Nothing is imputed. A county with no multi-year stations is absent from the ranking — not estimated from its neighbors, not smoothed, not filled. Every published figure is also stamped with the most recent count year behind it, because a corridor ranking measured through 2024 and one measured through 2017 are different objects and should never be presented as the same one. Among the states above, the Utah panel runs through 2024–25, Arizona and Maryland through 2023, North Carolina through 2022, and Mississippi through 2021.
Where this goes next
Coverage follows what each transportation department publishes. Thirteen states currently expose a multi-year count layer, which is what makes a growth rate computable; the rest publish a current-year snapshot, and three publish counts without year attribution at all. South Carolina is the clearest near-term opportunity — 9,483 stations already collected, all carrying 2025 counts, one multi-year source away from joining the ranking. County-level attribution follows the same path: every station in the panel carries coordinates, so extending county grain beyond North Carolina to the other twelve states is an attribution pass over data already in hand, and it is underway.
The larger point is what this data is. Vehicle counts sit in the same place as deeds, permits, and parcel records: published by a public agency, free to anyone willing to go get them from several thousand separate sources, and almost never assembled into something you can rank. We went and got them.
Methodology and data: PropRaven Corridor Momentum (v0) measures annualized corridor traffic growth as the ordinary-least-squares slope of log annual average daily traffic on calendar year, over a trailing ten-year window, at each state transportation department count station PropRaven collects. Station rates roll up to county and state by a mean weighted on the log of the latest count, reported alongside the unweighted median and the share of stations growing, then standardized to a within-grain percentile. Guards: count year 1985 to present; minimum two dated observations spanning two years; minimum latest count of 200 vehicles per day; growth beyond ±50% per year dropped; remainder winsorized at the 1st and 99th percentiles within state. Of 1,041,762 collected stations across 50 states, 147,600 carry a multi-year panel and clear these guards. Rankings are relative within their grain and are stamped with the most recent count year behind each geography. Nothing is imputed; geographies below threshold are absent rather than estimated. Figures are as of 2026-08-01 and are restated as new count years publish.
The parcel, deed, permit, and traffic records behind these numbers — 238M+ parcels and counting — are searchable at propraven.com.