Commuting Mode Share By Census Tract Data: Dataset Structure and Field Coverage

Commuting mode share by census tract is the split of a tract's workers 16 years and over by their means of transportation to work, and the Census Bureau publishes it as Table B08301 (Means of Transportation to Work) inside the American Community Survey. The table divides workers into drove alone, carpool, transit, walk, bicycle, motorcycle, taxicab and worked from home, with the transit cell broken further into bus, subway or elevated, long-distance train, light rail and ferryboat subcategories. Roughly 25 variables cover about 840 county equivalents and approximately 85,000 tract geographies in the 5-year release.

The ACS has run annually since 2005, and the releases currently documented on the program page are the 2024 1-year ACS and the 2020-2024 5-year ACS with new PUMS files. Because 1-year estimates stop at populations of 65,000+, the 5-year product is the one that reaches every tract - and every block group beneath it.

Exactly one record among the 14 passenger-ground-transportation datasets Datadory catalogs reports it: American Community Survey - Journey to Work / Commuting Data, published by the U.S. Census Bureau, scoring 9 out of 10, commercial delivery terms as a U.S. government work, and reachable through the Census Data API, data.census.gov and bulk summary files.

Everything else in the slice sits higher up the geography stack:

• National Transit Database (NTD) Monthly & Annual Ridership Datasets - about 370,000 agency x mode x month rows across 834 NTD IDs since January 2002, identified by agency, urbanized area, state and FTA region, with no tract field anywhere.
• NYC TLC Trip Record Data (Yellow/Green Taxi & FHV/HVFHV Uber/Lyft) - one row per trip since January 2009, but locations coded as 258 TLC Taxi Zones rather than census tracts.
• Eurostat Passenger Transport Statistics (tran_hv_psmod / rail_pa) - harmonized inland modal split for 37 European geographies annually from 1990 to 2024, national level only.
• National Household Travel Survey (NHTS) - person- and trip-level microdata whose household file carries Census division, region, CBSA and state FIPS geography, stopping well short of tracts.

The ACS is also the only source here with a tract option at all: its 5-year estimates cover every geography down to block groups, while the 1-year release serves geographies of 65,000+ population. That makes the 5-year vintage the default for anyone comparing a downtown tract against a suburban one.

Six verified variables carry a complete tract-level mode-share table; the remaining cells - bicycle, motorcycle, taxicab and worked from home - sit in the same B08301 group. Each variable's API label mirrors the Census hierarchy, so Estimate!!Total:!!Car, truck, or van:!!Drove alone is the raw numerator for the drove-alone column.

Three companion tables round out a tract profile once the shares exist: B08303 bands travel time to work into minute intervals, B08134 cross-tabulates vehicles available against means of transportation, and the B07404/B07410 series track residence-year-to-residence-year flows. Together they let one tract query answer not just how workers travel but how long they travel and how many households hold no vehicle at all - the difference between a walkable core and a car-dependent periphery usually showing up in all three at once.

The same group also fixes two habits worth breaking early. First, never rank tracts on raw counts: a tract with 900 drove-alone workers can be less car-dependent than one with 400 out of 450 total. Second, keep the transit cell whole when benchmarking against ridership data - its bus, subway, long-distance train, light rail and ferryboat splits answer network questions, not mode-share questions.

On the supply side, the NTD's roughly 370,000 monthly rows describe the service levels a tract's transit share rests on - though its identifiers end at urbanized area, state and FTA region. And when a catchment question needs home-to-work pairs rather than home-side shares, the county-to-county commuting flows tables in the same ACS release pair origin and destination counties.