Cradle & Quill makes a baby-name discovery app: expecting parents browse names, save favourites and share shortlists. The product team wants to add features built on real naming data, such as “rising names”, “names that work for any gender” and regional picks, and has asked you to look at the U.S. Social Security Administration’s baby-name records first.
Simple lookups (“the most popular name”) are easy. The hard part is choosing definitions that hold up: what counts as a rising name, how to measure how gender-neutral a name is, and what the data cannot tell you. Those choices end up in front of users, so they need to be defensible.
The data
names_by_state— one row per state, sex, year of birth and name, withcount, the number of babies of that sex born in that state that year who received the name. Source: U.S. Social Security Administration, based on Social Security card applications, by state of birth, from 1910 onward.- SSA omits a name in a state-year when fewer than 5 babies got it, so rare names, and the early years of names that later became popular, are missing rather than zero. Summing across states therefore gives less than the true national total.
Your task
- Describe the structure of the data and its limitations. What distortions could the collection process, the 5-baby threshold, spelling variants or the state-of-birth breakdown introduce, and how would each affect the measures below? Show evidence from the table where you can.
- Define a measure of gender ambiguity for a name and defend it. How does it handle names that are rare, or common for one sex and rare for the other? Use it to compare the most ambiguous names of 1945 and 2013, and explain what changed.
- Define what it means for a name’s popularity to have increased or decreased since 1980.
- Decide between raw counts and shares of births, and how to treat names absent (or below the threshold) in 1980 or in the latest year.
- Report the largest risers and fallers under your definition, and show how sensitive the ranking is to those choices.
- Which names might have changed even more than the data shows, and how could you find them (for example, by grouping spelling variants)?
- Find one insight in this data that Cradle & Quill could act on, such as how names spread between states, how quickly a name rises and fades, or whether next year’s risers can be predicted. Explain how you would check that it holds up (for example, by testing it on past years), and what other public data you would combine with it.
What to deliver
A short written report (3–5 pages with charts) for the Cradle & Quill product team: your definitions, findings and the feature you would build first. Show the queries or code behind each number so the work can be reproduced.
All company names in this question are fictional.