This module and the associated data download provide the total number of individuals who worked from home (WFH) on the modeled day that meet your Study's filter criteria. This data is generated by Replica based on observed mobility patterns in the modeled season; see a description of our methodology below.
The following categories are used for our WFH data:
Worked in-person: Employed residents who worked in-person at their place of employment or a job site on the modeled Thursday or Saturday.
Worked from home: Employed residents who worked from home on the modeled Thursday or Saturday and did not travel to their place of employment or job site.
Employed but did not work on this modeled day: Employed residents who did not travel to work on the modeled Thursday or Saturday (e.g. because they are employed with part-time work, or took a day off for vacation or illness).
Unemployed, under 16, or not in labor force: Residents who do not have a Working-From-Home status because they are not employed: they are either unemployed, under 16, or not in the labor force.
Note: to filter residents by their Census-defined employment status (e.g. "Employed" or "Not in the labor force"), use the "Employment Status" module.
WFH Methodology
Starting with the Spring 2021 season, we model the number of people that work from home based on two data types: mobile location data, and surveys conducted by the Bureau of Labor Statistics (BLS). The BLS survey provides a lagging monthly estimate of the nationwide share of employed people working from home by industry of employment (by NAICS code). We use mobile location data – specifically, the number of hours observed devices spend at home and work locations – to:
1. Project the BLS estimate forward to the current time period, and
2. Model spatial variations in remote working for each Census Public Use Microdata Area (PUMA), the smallest geographic unit for which the Census reports detailed employment data (Public Use Microdata Sample or “PUMS” data is delivered for each PUMA). Each PUMA consists of 100,000-200,000 people.
Using the above estimate, we randomly assign each person in our synthetic population with the appropriate home location and industry of employment to a “Work-From-Home” persona. These people will spend a period of time working from home, and will not make a commuting trip to their place of employment. They may still make trips for other purposes throughout the day.