We now hold something that didn’t exist before: a consistent, sourced, comparably-scored profile of 1,499 American cities and towns — the mapped bike network in each, the share of commuters who ride, the terrain, and the shape of the riding year. Each fact carries its own source and date. Nothing in the set is estimated or hand-entered.
Any one city profile answers a personal question: is my town good for riding? Read together, they answer a civic one: what actually stands between an ordinary American town and everyday cycling?
The short answer is that it is usually not the weather, and it is usually not the number of miles.
Miles don’t make riders
If bike lanes produced riders in any simple way, mileage and ridership would move together. Across the 1,490 towns where we hold both figures, they barely do: the correlation between mapped network miles and bike-to-work share is r = 0.3 — positive, real, and weak. Mileage explains a small part of the story and leaves most of it unexplained.
That number is worth sitting with, because it cuts against the way both sides of a council argument usually talk. It does not say infrastructure doesn’t work. It says miles are the wrong unit. A town can hold two hundred miles of path and no way to cross the arterial between you and the grocery store.
Where the bones outrun the habit
64 towns sit in the top quarter of the Compass for network mileage and the bottom quarter for people riding to work. They have built something real, and the riding hasn’t followed.
| Town | Mapped network | Bike-to-work share |
|---|---|---|
| Woodbury, MN | 252.0 mi | 0.1% |
| Laguna Hills, CA | 246.3 mi | 0.1% |
| Plymouth, MN | 237.7 mi | rounds to 0% |
| Brooklyn Park, MN | 213.5 mi | 0.1% |
| Eagan, MN | 204.1 mi | 0.1% |
| Fountain Valley, CA | 197.5 mi | 0.1% |
The list has a shape. It is dominated by prosperous suburbs, the Twin Cities ring above all, and suburban Southern California, with generous, enviable trail systems. These towns invested heavily; they simply built beautiful recreational networks that most people drive to reach.
That is the lesson, and it isn’t a scolding one: a trail that loops through a park does a different job than a route to the school. In these towns, another mile of path is rarely what’s missing. The gaps are at the crossings, the connections, the last quarter mile into the places people actually go, plus the calm, ordinary streets in between. That last one is the dimension where nearly every American town in our set scores lowest.
Where the habit outruns the bones
The mirror image is more encouraging. 48 towns are in the bottom quarter for network and the top quarter for riding. People ride there anyway.
| Town | Mapped network | Bike-to-work share |
|---|---|---|
| Galveston, TX | 8.6 mi | 2.0% |
| Binghamton, NY | 9.4 mi | 1.6% |
| Savannah, GA | 8.5 mi | 1.5% |
| Port Chester, NY | 3.8 mi | 1.5% |
| Bangor, ME | 4.2 mi | 1.3% |
What these places share is shortness. They are compact, older, walkable-scaled towns, several of them college towns, where the trip you need to make is a mile rather than six. The short distances do what bike lanes alone can’t.
The weather excuse doesn’t survive the data
“It’s the climate” is the most common explanation offered for why a town doesn’t ride. Our season data says otherwise.
19 towns enjoy ten or more good riding months a year — the climate you would design if you could — and still sit in the bottom quarter for riding. Nearly all of them are in Southern California, and several of them, like Laguna Hills and Fountain Valley, also hold big networks. Twelve good months, two hundred miles of path, and almost nobody riding to work.
Meanwhile, the median town in the whole set gets seven good months. That is a riding year most of the world would take.
The state of the American bike network, in one table
Across all 1,499 towns, here is where the six Compass dimensions actually land. The pattern is stark and, we think, the most useful thing in this report:
| Dimension | Median level | What it means |
|---|---|---|
| All-Season | 50 — Solid | The riding year mostly works. Climate is the one dimension where the median American town is already fine. |
| Room to Roam | 36 — Growing | How far you can get. |
| Connected | 34 — Growing | Whether the pieces join into routes. |
| Welcoming | 31 — Growing | Whether a newcomer or a nervous rider could start. |
| Calm | 30 — Growing | How much riding is away from fast traffic. The lowest of the six. |
| Car-Light | 22 — Growing | How well the bike replaces a car trip today. |
One dimension out of six is in decent shape, and it is the one nobody built: the weather. Every dimension a town can actually decide — connection, calm, range, welcome — sits in the Growing band for the median American town. Calm is the floor. Not miles. Not sunshine. The plain, unglamorous business of separating a rider from fast traffic.
What we can’t tell you
The caveats are the reason to trust the rest, so here they are without being asked:
- OpenStreetMap mileage is a floor, not a survey. A low figure can mean a thin network or a thinly mapped one. Volunteer mapping is uneven, and some of the towns in the “habit outruns the bones” list are better provided-for than their mileage suggests.
- The Census bike-to-work share counts commutes only. It misses the school run, the grocery trip, the ride to the trailhead, and everyone who doesn’t commute at all. It undercounts real riding everywhere — which is exactly why it is still useful as a comparison between places measured the same way.
- Small-town ACS estimates carry wide margins of error. Treat any single town’s number as indicative, not exact. The patterns across 1,499 of them are far sturdier than any one row.
- These are associations, not causes. Nothing here proves that connecting a network produces riders in a given town. It shows what the built environment and the riding look like when you line up 1,499 of them side by side.
Re-run it yourself
Every figure in this piece is computed from the published city files by scripts/compass-report.mjs, which reads the same sourced data any reader can see on any city page. Re-run it and the numbers move with the data; the full output is committed alongside it. We would rather show the arithmetic than ask to be believed.
If this is your town on one of these lists, don’t take it as a scolding. Take it to the next meeting, and ask the one question the data really supports: what does this connect? Our guides on reading your city’s bike plan, making the case, and two minutes at city hall are the practical follow-ons.
Whenever you can, ride.