of commuters traveled 45 minutes or longer one way. Rank 29 of 35; 90% margin of error ±0.7 percentage points.
Before the Workday
Compare how long workers travel in the 35 largest U.S. metro areas, see how commute modes differ, and translate a typical round trip into a transparent annual-hours scenario.
Start here
A quick guide to this quest.
A commute is time spent getting to work before the paid day begins. This project compares the 35 largest U.S. metro areas using 2024 survey estimates, while keeping the uncertainty and limits of those estimates visible.
Start with Indianapolis
Use the opening button to compare Indianapolis with New York. Notice how the 45-minute-plus share and the full time distribution differ.
Choose two metro areas
Change either selection or choose a dot in the comparison field. Every chart and scenario updates to keep the same pair in view.
Select a time band
Choose any block in the travel-time rows to compare that exact published band across both metro areas.
Change the weekly scenario
Move the commute-days slider to see how the reported average one-way trip adds up under a simple 50-workweek model.
- Metropolitan area
- A Census-defined region built around an urban center and nearby communities linked to it by work and economic activity.
- Commuter
- Here, a worker age 16 or older who traveled to work; people who worked from home are excluded from travel-time results.
- Percentage point
- The direct difference between two percentages; 20% minus 15% is a difference of 5 percentage points.
- 90% margin of error
- A survey uncertainty range: under repeated sampling, about 90% of intervals built this way would contain the population value.
- ACS
- The American Community Survey, an ongoing Census Bureau sample survey that produces annual estimates about U.S. communities.
2024 American Community Survey · 35 largest U.S. metro areas
Before the
workday.
A commute is time that sits outside the paid day. This project compares how long workers travel, how they travel, and how much a typical round trip can add up across a year.
Begin here · 01
Start with Indianapolis
Compare its commute with New York, then choose any two metro areas.
What is measured? One-way travel time for workers age 16 and older who did not work from home.
Why 45 minutes? It is the first published Census time band beyond 40–44 minutes, so no individual trips were reclassified.
What is uncertain? These are survey estimates, not exact counts. Error bars show the reported 90% margin of error.
Choose · 02
Choose two metro areas
Both choices update every view below. The first metro is coral; the comparison metro is lime.
of commuters traveled 45 minutes or longer one way. Rank 1 of 35; 90% margin of error ±0.4 percentage points.
New York-Newark-Jersey City has the larger estimated share.
Scan · 03
Where long commutes are more common
First metro Comparison
Move right for a larger work-from-home share and up for a larger share traveling at least 45 minutes. Select any dot to make it the first metro. Vertical lines show the 90% margin of error for the long-commute share.
Read the ranked values as a table
| Rank | Metro area | 45+ minutes | 90% margin | Worked from home |
|---|---|---|---|---|
| 1 | New York-Newark-Jersey City, NY-NJ | 33.5% | ±0.4 points | 12.9% |
| 2 | Washington-Arlington-Alexandria, DC-VA-MD-WV | 28.9% | ±0.6 points | 19.9% |
| 3 | Riverside-San Bernardino-Ontario, CA | 27.4% | ±0.7 points | 10.9% |
| 4 | San Francisco-Oakland-Fremont, CA | 27% | ±0.7 points | 18.9% |
| 5 | Atlanta-Sandy Springs-Roswell, GA | 26.1% | ±0.6 points | 18.8% |
| 6 | Boston-Cambridge-Newton, MA-NH | 26.1% | ±0.6 points | 16.2% |
| 7 | Houston-Pasadena-The Woodlands, TX | 25% | ±0.6 points | 12.1% |
| 8 | Chicago-Naperville-Elgin, IL-IN | 24.3% | ±0.5 points | 14.9% |
| 9 | Los Angeles-Long Beach-Anaheim, CA | 22.7% | ±0.3 points | 14.9% |
| 10 | Baltimore-Columbia-Towson, MD | 22.6% | ±0.8 points | 15.1% |
| 11 | Miami-Fort Lauderdale-West Palm Beach, FL | 21.7% | ±0.5 points | 14.3% |
| 12 | Tampa-St. Petersburg-Clearwater, FL | 21.7% | ±0.9 points | 19.3% |
| 13 | Seattle-Tacoma-Bellevue, WA | 21.6% | ±0.6 points | 18.5% |
| 14 | Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 21.3% | ±0.6 points | 15.4% |
| 15 | Nashville-Davidson--Murfreesboro--Franklin, TN | 20.7% | ±0.9 points | 16.8% |
| 16 | Dallas-Fort Worth-Arlington, TX | 19.8% | ±0.5 points | 15.6% |
| 17 | Austin-Round Rock-San Marcos, TX | 18.9% | ±0.9 points | 23.2% |
| 18 | Orlando-Kissimmee-Sanford, FL | 18.5% | ±0.9 points | 16.8% |
| 19 | Charlotte-Concord-Gastonia, NC-SC | 17.3% | ±0.7 points | 19% |
| 20 | Phoenix-Mesa-Chandler, AZ | 17.1% | ±0.6 points | 17.8% |
| 21 | Denver-Aurora-Centennial, CO | 16.4% | ±0.6 points | 22.6% |
| 22 | Pittsburgh, PA | 16.3% | ±0.7 points | 15.7% |
| 23 | San Antonio-New Braunfels, TX | 16.1% | ±0.8 points | 13.4% |
| 24 | Sacramento-Roseville-Folsom, CA | 15.9% | ±0.8 points | 16.8% |
| 25 | Detroit-Warren-Dearborn, MI | 15% | ±0.5 points | 12.5% |
| 26 | San Diego-Chula Vista-Carlsbad, CA | 14.2% | ±0.6 points | 16.1% |
| 27 | Portland-Vancouver-Hillsboro, OR-WA | 14.1% | ±0.7 points | 20% |
| 28 | St. Louis, MO-IL | 13.2% | ±0.6 points | 13.3% |
| 29 | Indianapolis-Carmel-Greenwood, IN | 13.1% | ±0.7 points | 13.6% |
| 30 | Cincinnati, OH-KY-IN | 12% | ±0.6 points | 12.7% |
| 31 | Las Vegas-Henderson-North Las Vegas, NV | 10.9% | ±0.8 points | 11.4% |
| 32 | Minneapolis-St. Paul-Bloomington, MN-WI | 10.7% | ±0.5 points | 17.4% |
| 33 | Columbus, OH | 10.7% | ±0.6 points | 16.9% |
| 34 | Kansas City, MO-KS | 10.1% | ±0.6 points | 14.7% |
| 35 | Cleveland, OH | 9.5% | ±0.5 points | 13.5% |
Compare · 04
How one-way travel time is distributed
Each row represents 100 commuters who traveled to work. Wider blocks mean more people fell in that time band. Select a block to compare one band precisely; the same selection applies to both rows.
Indianapolis-Carmel-GreenwoodAverage one-way trip: 25.6 minutes
New York-Newark-Jersey CityAverage one-way trip: 36.5 minutes
Selected time45 minutes or longer
Indianapolis-Carmel-Greenwood13.1%
New York-Newark-Jersey City33.5%
20.4 percentage points apart in this selected band.
Explain · 05
How workers usually got to work
Each horizontal bar is a share of all workers age 16 and older, including people who worked from home. Compare the labels and values; color is only a secondary cue.
Indianapolis-Carmel-Greenwood, IN
2.2 million residents · 960,422 commuters with a reported travel time
- Drove alone74.6%Car, truck, or van with one occupant
- Carpooled8.3%Two or more people sharing a car, truck, or van
- Public transportation0.6%Bus, rail, streetcar, subway, or ferry
- Walked or bicycled1.6%The two people-powered categories combined
- Worked from home13.6%No commute trip reported
- Other travel1.4%Taxi or ride-hailing, motorcycle, and other means
New York-Newark-Jersey City, NY-NJ
19.9 million residents · 8,545,850 commuters with a reported travel time
- Drove alone44.5%Car, truck, or van with one occupant
- Carpooled6.6%Two or more people sharing a car, truck, or van
- Public transportation27.2%Bus, rail, streetcar, subway, or ferry
- Walked or bicycled6.3%The two people-powered categories combined
- Worked from home12.9%No commute trip reported
- Other travel2.4%Taxi or ride-hailing, motorcycle, and other means
Translate · 06
Put the average trip on a yearly clock
This scenario doubles the reported one-way average for a round trip, then multiplies it by your chosen days per week and 50 workweeks. It is an illustration—not a prediction of any person’s schedule.
213 hours
about 27 eight-hour days in this 5-day scenario
304 hours
about 38 eight-hour days in this 5-day scenario
Changing the slider does not change the Census estimate. It only changes this transparent annualization model.
Why this exists
The question.
The share of workers traveling at least 45 minutes varies substantially among large metro areas, and showing work-from-home shares, travel modes, uncertainty, and the full time distribution prevents a single ranking from becoming a misleading explanation.
How to begin.
Start with Indianapolis and New York, then choose any two metro areas to update the comparison field, time distribution, travel modes, and annual-hours scenario.
Method and limits
How the project was built.
I downloaded the official 2024 American Community Survey one-year table-based Summary File and joined metropolitan-area rows from four detailed tables: B01003 for resident population, B08013 for aggregate travel minutes, B08301 for means of transportation to work, and B08303 for one-way travel-time bands. I selected the 35 largest metropolitan areas by the B01003 population estimate, divided category estimates by their published table universes, and rounded displayed shares to one decimal place. The mean travel time divides aggregate travel minutes by workers with a reported trip. The margin for the 45-minute-plus share combines the three source count margins in quadrature and applies the Census Bureau's approximation for a percentage whose numerator is a subset of its denominator. ACS values are sample-based estimates with sampling and nonsampling error. Metro boundaries, population size, work location, land use, housing, and transport access differ, so the project describes patterns without assigning causes or predicting any person's commute.
What this demonstrates
- Data engineering
- Statistical reasoning
- Product thinking
- Visual design
- Editorial storytelling
- Interaction design
- Accessibility
Takeaways
- In the fixed 2024 estimates, 13.1% of Indianapolis-area commuters traveled at least 45 minutes one way, compared with 33.5% in the New York area. The corresponding approximate 90% margins of error are 0.7 and 0.4 percentage points.
- Indianapolis ranks 29th and New York ranks 1st by estimated 45-minute-plus share among these 35 large metro areas, but a rank alone cannot explain the difference: work-from-home shares, travel modes, metro structure, and other unmeasured conditions must be read alongside it.
- The annual-hours view is a transparent scenario based on average one-way time, two trips per day, a chosen one-to-five commute days per week, and 50 workweeks. It is not a Census estimate of annual time and is not a personal forecast.