SparkyData

NYC Remote Work and Commuting: Census Microdata

Public-use microdata from the article “Who still goes to the office”

Not synthetic. Disclosure-protected survey records published by a government statistical agency, cut to the people or households we looked at.

Records
Public-use microdata
Place
New York City
Version
1.12,
Licence
CC BY 4.0; code MIT
Built for
Who still goes to the office (version 1.12)

Working from home has settled rather than faded, and it has settled on the better paid. In 2026 about a quarter of New Yorkers at work spend some or all of the week working from home, one in ten all of it, as in 2025. Before the pandemic the share was the same at every pay level; now it climbs with pay. Seven in eight still travel to work, and the commute that shrank is the one to Manhattan.

The article's own data file, microdata cut to what it analyses, and the script that computes every figure in it. Computed from: ACS 2015–2024 one-year PUMS, NYC and metro-area resident workers, with the 2019 and 2024 files for the comparisons; CPS basic monthly files June 2024 to August 2026 for the telework question; MTA daily ridership to 30 September 2026.

Files

FileRowsSizeWhat it holds
data.csv 66,499 11.4 MB The microdata the article's figures are computed from.
analysis.py 56 KB The script that computes every figure in the article (MIT).

Columns

data.csv: The data (22 columns)

row_id, year, weight, borough, worked_from_home, mode, mode_group, one_way_minutes, workplace_manhattan, physically_commutes_to_manhattan, own_earnings_2024, earnings_band, occupation_group, industry_sector, office_based_industry, class_of_worker, usual_hours, race_ethnicity, foreign_born, bachelors_or_higher, age, sex. The article and its script define each one.

Sources

Licence and credit

The survey records are the publishing agency's, redistributed as published; the columns we derived are CC BY 4.0; the script is MIT. Public-use microdata are disclosure-protected records derived from survey responses, published by the issuing agency so that no respondent can be identified; synthetic records are generated by SparkyData and correspond to no real person, household or address. Neither may be used to identify anyone.

Credit it as:

SparkyData (2026). NYC Remote Work and Commuting: Census Microdata (version 1.12). Data CC BY 4.0.
https://sparkydata.dev/data/nyc-remote-work-and-commuting-census-microdata

The article gives the charts, the uncertainty of each comparison and a changelog. Corrections and questions: sparkytech.dev@gmail.com, or see corrections.