Fake data for tests
Test suites and design comps need rows that look like production without being production. A repeatable seed is more useful than a purely random draw: the same seed and schema give you the same 25 people every time, so a failing test stays failing until you change the code.
The names, cities and phone numbers on this page are invented and are labelled as example.com addresses. Do not use them as if they were real people. If you need statistically fair populations or locale-accurate national IDs, a specialised library is a better fit than a 100-row notebook.