How we estimate rush-hour congestion

The traffic layer is an estimate. For each road, we work out how full it gets in a typical weekday rush hour, using traffic counts the states report to the Federal Highway Administration. Neighborhoods next to roads that run near or over their capacity are marked as congested. Nothing on this layer is a live or recorded traffic speed.

What does the layer show?

Every place on the map falls into one of four levels:

• Free flow. Nearby roads have room to spare at peak.
• Moderate. A nearby road is filling up. Expect slowdowns.
• Heavy. A nearby road is close to full. Stop-and-go is likely.
• Severe. More traffic wants a nearby road than it can carry.

You pick how much you can live with. “Avoid heavy congestion” hides the heavy and severe areas. “Avoid all rush-hour congestion” hides anything that is not free flow. What is left shows in yellow.

Where does the data come from?

From the Highway Performance Monitoring System, usually called HPMS. Every state transportation department reports its roads to the Federal Highway Administration each year. For each stretch of road, that includes how many vehicles use it on an average day, how many lanes it has, and what kind of road it is.

The current layer uses the 2024 data year, which the U.S. Department of Transportation published in 2026. You can find it as Highway Performance Monitoring System (HPMS) 2024, part of the National Transportation Atlas Database. It is a U.S. government work in the public domain. FHWA asks that it be cited as the source, so here it is: Federal Highway Administration, Highway Performance Monitoring System, distributed by the Bureau of Transportation Statistics.

FHWA puts out a new year once a year, and we rebuild the layer when it does.

How do you turn traffic counts into congestion?

We use the method FHWA itself uses for national planning. It is set out in the HPMS Field Manual, in its appendix on estimating highway capacity, and it builds on the Highway Capacity Manual.

It comes down to two numbers for each road.

The first is how much traffic arrives in the busiest hour. A road’s daily count is spread over the day, so we take the share that falls in the peak hour, and the share heading the busier way.

The second is how much the road can carry in that hour. That depends on the kind of road. A freeway lane carries far more than a lane on a street with traffic lights, because the lights stop the flow part of the time. Trucks take up more room than cars. A two-lane country road works differently again, because drivers share it with oncoming traffic.

Divide the first by the second and you have how full the road is. The fuller it is, the higher the level.

Where a state reports the details for a road, such as its peak-hour share or how long its lights stay green, we use them. Where it does not, we use typical values for that kind of road: FHWA’s planning defaults, or, for traffic lights, the timing states most commonly report for similar roads.

Why is a whole neighborhood marked, not just the road?

Because congestion is felt nearby, not only on the road itself. If the arterial at the end of your street is jammed every morning, that is your commute too. So each congested road marks a band a few hundred meters wide on either side.

Where bands from different roads overlap, the worse road decides. A quiet street next to a jammed freeway reads as the freeway.

What does it get wrong?

Several things, and they are worth knowing before you rely on it.

• Traffic lights are estimated. For most city streets we assume typical signal timing. A badly timed corridor can be worse than it looks here, and a well coordinated one can move better.
• It shows a typical weekday. Construction, crashes, school terms, holidays and tourist seasons do not appear.
• Small local streets are mostly missing. States are only required to count traffic on the larger roads, so most residential streets have no count. With no count, they read as free flow.
• It is one to two years old by the time you see it. A new subdivision or a new interchange takes a while to show up.
• The peak hour it uses is a busy one. Planners size roads for a heavier hour than the average weekday, so this layer leans toward showing more congestion, not less.

What don’t you use?

This layer uses no commercial or licensed traffic feed. We do not use any navigation app’s traffic data, live or historical, and we do not sample any traffic map. Every number behind it comes from the public FHWA data above.

Can I check it myself?

Yes. The HPMS file linked above is free to download. Open it in any mapping program, click a road near the address you care about, and look at its average daily traffic, its number of through lanes and its road type. Those are the main things the estimate is built from.

If a road looks wrong to you, the count is the first thing to check. Counts on some roads are years old even in a new release, because states do not count every road every year.

The traffic congestion layer is on the map if you’d like to look at an area you know.