×
THE DATA, IN PLAIN ENGLISH
Every diary since 2003.
258,954 complete diaries. 4,994,172 activity records. Twenty-three years of everyday life.
What do these numbers mean?
Estimates for the U.S. civilian, noninstitutional population age 15 and older at the
selected local clock time, or for the people matching your filters. Counts represent the
average population across 2003–2025, not today’s population. “Your time” uses your
device’s clock. This is not a live or time-zone-synchronized national count, or a
prediction for a specific date.
Combining years gives more diaries but blends real changes over time, such as the
pandemic. The “Across the years” chart shows each year separately so you can see those
changes.
How are years and activities combined?
Each person’s final survey weight contributes during their reported activity interval:
TUFNWGTP (the 2006 method, comparable across all years) for every year except 2020, which
uses TU20FWGT, as BLS directs. Counts are divided by the number of days the 23 years
represent. Data collection paused from March 18 to May 9, 2020, so 2020 represents 313
days. After the October 1–November 12, 2025 shutdown, BLS weighted the remaining
fourth-quarter diaries to represent the whole quarter. A diary begins at 4 AM; its
after-midnight portion keeps the starting day’s weekday/weekend assignment.
The overview partitions all activities into 17 groups. Detailed view uses every pooled
six-digit activity code. BLS harmonized a few codes so they mean the same thing across all
years; their descriptions list the original codes they combine. Reported drivers (TEWHERE
= 12) are assigned to driving first, so detailed rows sum to their overview parent.
Filters, places, and company
People filters (sex, age, work status, children at home) use each respondent’s answers on
their diary day. Children means household children under 18, grouped by the youngest
child’s age. One 2010 respondent with no full- or part-time status is counted as full
time. Advanced filters add demographics, education, work, earnings, household, and diary
day from the ATUS files, and family income, marital status, citizenship, veteran status,
disability, region, and state from the household’s final Current Population Survey
interview, two to five months before the diary day. Every filter accepts several options
at once. Each menu notes when a question was first asked or how it is measured.
Activity-context filters keep only activities recorded at that place or with that company.
While one is on, the summary shows everyone in that context (for example, everyone at a
restaurant, whatever they are doing) instead of Sleeping and Awake, and every number on
the page follows the same filters. BLS doesn’t record location or company for sleeping,
grooming, or private activities, and didn’t record company while working before 2010.
Someone can be with several groups at once, so company shares can add up to more than
100%.
Sleeping, awake, and sample size
Sleeping includes sleep and naps (010101 and 010199), excluding reported sleeplessness.
Awake means everyone not classified as sleeping, including unclassified time; it is not a
direct measurement of consciousness. Driving includes car, truck, and motorcycle drivers,
not passengers. Primary activities do not capture all multitasking.
Every selected activity shows its actual unweighted number of diaries at that minute.
Fewer than 30 observations is flagged as a small sample; this is a caution rule, not an
official BLS publication threshold. Narrow filters can leave few diaries. Zero
observations does not establish that nobody was doing the activity. Confidence intervals
have not been computed.
Sources and validation
BLS 2003–2025 multi-year microdata
·
Pooling guidance
·
Weighting methods
·
2003–2025 activity lexicon
Every diary covers 1,440 contiguous minutes. The data reproduces all 96 hourly sleeping
values in BLS Table A-3 for 2021–2024 and the published 2025 average of 9.03 hours of
sleep. Detailed activities sum to their parent at every minute, and filtered estimates are
checked against an independent pass over the raw BLS files. This is an independent
visualization, not an official BLS publication.