Distances and Travel Times from Census Tracts to Nearby Medicaid offices in the United States

DOI: 10.5281/zenodo.20684650Previous version: 10.5281/zenodo.18779578License: CC BY 4.0 Overview This dataset provides census tract–level measures of geographic accessibility to official state-operated Medicaid offices across the United States. It includes multi-modal travel time and distance estimates from census tract centroids to the nearest Medicaid office, along with tract-level demographic and socioeconomic characteristics. The dataset covers all census tracts in the 50 states and the District of Columbia (DC) — 83,998 tracts in total. It was constructed to examine spatial inequalities in administrative access to Medicaid services and to support research on geographic barriers to public insurance enrollment. Data Sources Medicaid Office Locations Medicaid office locations were obtained from a publicly available national inventory of official state-operated Medicaid offices providing in-person enrollment assistance, identified between 2021 and 2022 and updated through late 2023 (Shafer et al., 2024; https://doi.org/10.1016/j.dib.2024.110068). Non-state-operated entities (e.g., private providers, FQHCs) were excluded to ensure coverage reflects official state administrative locations only. Census Tract Characteristics Neighborhood demographic and socioeconomic characteristics were drawn from the 2016–2020 American Community Survey (ACS) 5-year estimates. Census tract polygons and centroid coordinates were derived from U.S. Census Bureau TIGER/Line shapefiles, standardized to the WGS84 coordinate reference system. Travel Time Estimation Travel times were estimated using the Google Maps Distance Matrix API, which integrates current road networks, speed limits, transit timetables, and pedestrian/bicycling infrastructure to return realistic, mode-specific travel times. Each value represents the minimum observed travel time from a census tract centroid to the nearest Medicaid office. Travel time queries differed by mode: Walking and bicycling: Queries were conducted during off-peak weekday hours — defined as after 20:00 and before 06:00 — to minimize variability from traffic conditions. Walking and bicycling travel times are treated as time-invariant (i.e., not meaningfully affected by time of day), so off-peak estimates are considered representative. Multiple queries were submitted across weekdays; the shortest observed travel time and distance were retained. Driving: Queries were conducted during two time windows: (1) off-peak weekday hours (after 20:00 and before 06:00) and (2) weekday mornings at 9:00 AM, which reflects more realistic conditions during typical Medicaid office hours. The reported duration_driving is the average of these two estimates when both are available; if only one estimate is available, that value is used. Public transit: Queries were conducted at 9:00 AM on weekdays, which aligns with morning peak service schedules and is representative of transit availability during Medicaid office hours. The 9 AM window is standard in the transit accessibility literature for capturing realistic service levels. Where no feasible transit route exists, values are missing — this reflects structural transit inaccessibility, not a data error. All API queries specified travel mode and used metric units for distance and time. Non-parseable entries were set to missing prior to analysis. File Description File Description 06052026 analytical data.xlsx Main dataset for analysis. All variables listed below are included. Missing values are preserved as-is; no imputation has been applied. 06032026 Full Dataset_V5.xlsx Raw full dataset prior to analytical variable construction. Included for reproducibility. 06052026_data_cleaning.Rmd R Markdown file documenting the cleaning steps used to derive the analytical dataset from the raw full dataset. Variable Dictionary Identifiers Variable Description GEOID 11-digit census tract FIPS code origin Coordinates of the census tract centroid used as the trip origin destination Address or coordinates of the nearest Medicaid office state U.S. state abbreviation Travel Time and Distance — Primary Analytical Columns Variable Unit Description duration_driving minutes Driving travel time to the nearest Medicaid office. Derived as the average of duration_driving_offpeak and duration_driving_pred_9am_min when both are available; if only one is available, that value is used. distance_driving km Driving distance to the nearest Medicaid office. duration_walking minutes Walking travel time to the nearest Medicaid office, queried during off-peak weekday hours (after 20:00 and before 06:00). distance_walking km Walking distance to the nearest Medicaid office. duration_bicycling minutes Bicycling travel time to the nearest Medicaid office, queried during off-peak weekday hours (after 20:00 and before 06:00). distance_bicycling km Bicycling distance to the nearest Medicaid office. duration_transit minutes Public transit travel time to the nearest Medicaid office, queried at 9:00 AM on weekdays. Missing values (~40.6% of tracts) indicate that no feasible transit route exists and reflect structural transit inaccessibility. distance_transit km Public transit travel distance to the nearest Medicaid office, queried at 9:00 AM on weekdays. Missing where no route exists. Travel Time and Distance — Source Columns These columns preserve the raw inputs used to construct duration_driving, duration_transit, and distance_transit. Variable Unit Description duration_driving_offpeak minutes Driving travel time queried during off-peak weekday hours (after 20:00 and before 06:00). duration_driving_pred_9am_min minutes Driving travel time queried at 9:00 AM on weekdays. duration_transit_pred_9am_min minutes Public transit travel time queried at 9:00 AM on weekdays. This is the source for duration_transit. distance_transit_pred_9am_km km Public transit travel distance queried at 9:00 AM on weekdays. This is the source for distance_transit. Straight-Line Distance Variable Unit Description straight_line_km km Straight-line (Euclidean) distance from tract centroid to nearest Medicaid office. straight_line_m m Same, in meters. straight_line_mi miles Same, in miles. Tract-Level Sociodemographic Characteristics (ACS 2016–2020) All proportion variables are expressed as values between 0 and 1. Variable Description median_household_income Median household income, inflation-adjusted to 2020 USD. poverty_rate Proportion of the population below the federal poverty line. white_proportion Proportion identifying as non-Hispanic White. black_proportion Proportion identifying as non-Hispanic Black. asian_proportion Proportion identifying as non-Hispanic Asian. hispanic_proportion Proportion identifying as Hispanic or Latino. two_or_more_races_proportion Proportion identifying as two or more races. higher_education_proportion Proportion with a bachelor's degree or higher. foreign_born_proportion Proportion who are foreign-born. elderly_proportion Proportion of the population aged 65 and older. disability_proportion Proportion with any disability. cognitive_difficulty_proportion Proportion with a cognitive difficulty. male_proportion Proportion who are male. female_proportion Proportion who are female. unemployment_rate Civilian labor force unemployment rate. Missing Data Variable(s) Approx. Missing Interpretation duration_transit, distance_transit, duration_transit_pred_9am_min, distance_transit_pred_9am_km ~40.6% Structural — no feasible public transit route exists between the tract centroid and the nearest Medicaid office. Should not be imputed; treat as transit inaccessibility. duration_driving ~0.05% (41 rows) Both off-peak and 9 AM driving queries returned no result, likely due to extreme remoteness or API limitations. median_household_income ~1.4% Likely due to ACS reporting suppression in small or low-population tracts. All other sociodemographic variables < 1% Minimal and non-systematic. Walking, bicycling < 0.3% Minimal; likely due to API timeouts during data collection. Citation If you use this dataset, please cite: Nie, X., Huang, Z., & Yu, G. (2026). Distances and Travel Times from Census Tracts to Nearby Medicaid Offices in the United States. Zenodo. https://doi.org/10.5281/zenodo.20684650 For the Medicaid office location source, please also cite: Shafer, P. R., Palmer, M., Cho, A., Lynch, M., Louis, P., & Skinner, A. (2024). A dataset of geocoded Medicaid office locations in the United States. Data in Brief. https://doi.org/10.1016/j.dib.2024.110068