How a person is made
LifeSim does not copy or sample real people. It draws each detail from a table of rounded public figures, so one person is made up but a thousand of them add up to something that looks like the country.
The people on this site come from the original LifeSim generator, so every seed keeps the person it always had. This page describes the rebuilt generator, which is not serving people yet. The original follows the same idea, but its draws and tables differ in detail.
The order of the draws
Later steps depend on earlier ones. That is why an 80-year-old is most likely retired and a 22-year-old never has a doctorate.
Place
Pick a setting (urban, suburban or rural), then a place in it. Bigger places are more likely, but small towns still come up.
Race or origin
Draw from that place's population mix.
Age and birthday
Draw an adult age from the US age distribution. The birthday is fixed for the seed.
Sex and name
Draw sex, then a first name that was common in the birth decade, then a last name weighted by race or origin.
Body
Height and BMI from national body measurements by sex. Blood type and lactose tolerance vary by race or origin.
School, work and money
Education from national attainment, limited by age. Work status depends on age. Income depends on work status and education.
Household
Marriage and home ownership depend on age. A spouse is a second person drawn to match: same place and household, a close age.
Everything else
Phone, email, social media, pets, religion, personality and habits, each from its own table.
Seeds
A seed starts a random number generator. Each detail has its own stream from that seed, so a change to how pets are drawn never changes anyone's name or birthday. A seed is repeatable against one reference dataset; this build uses dev-2026.09.
Sources
Each table is a rounded version of the figure named here. The figures in this build have not yet been checked line by line against the published tables, so treat them as close, not exact.
| Detail | Weighted from |
|---|---|
| age | US Census Bureau, 2023 population estimates by age (adults 18+) |
| density | Pew Research Center, urban/suburban/rural self-description |
| body | CDC NHANES 2015-2018 anthropometric reference data |
| blood | American Red Cross, blood types by ethnicity |
| education | US Census Bureau, CPS educational attainment 2023 (age 25+) |
| income | US Census Bureau, median earnings by education 2023 |
| marriage | US Census Bureau, ACS marital status by age |
| housing | US Census Bureau, homeownership by age of householder |
| pets | APPA National Pet Owners Survey 2024 |
| veteran | US Department of Veterans Affairs, veteran population by period |
| social | Pew Research Center, social media use 2024 |
| vehicles | US Census Bureau, ACS vehicles available |
| phone | Pew Research Center, mobile fact sheet; StatCounter US OS share |
| religion | Pew Research Center, Religious Landscape Study 2023-24 |
| commute | US Census Bureau, ACS means of transportation and travel time |
What the data is not
- Not real people. Any match to a living person is chance.
- Not a statistical sample. The tables are rounded and simplified; do not use a batch as survey data.
- Not for contact. The emails are made up but use common mail domains, so one can belong to a real person. Do not send mail to them.
- US only, adults only. The sex field has two values because the v2 schema that clients depend on has two; see the field reference.