How Every Chart on This Site Is Computed (Full Methodology)

Ad transparency and reader trust both require the same thing: you should be able to reproduce every chart on this site from the raw files. Here are the exact rules. None of them are editorial opinions; all are plain calculations over public Social Security counts.

Trend (shown on every name page): take the most recent year the name appears in the data, call it L. Compare the total births in the five years L-4 through L against the five years L-9 through L-5. The trend is the percentage change. A name suppressed to zero in the earlier window has no trend — we show a dash rather than pretend a zero is real.

Rising and falling charts: names with at least 1,000 lifetime births, still active in the latest year, whose peak year is more than a decade in the past (we exclude names whose entire existence is the spike — a one-hit surge is not a trend), ranked by the trend above.

Evergreen: names on the national list for at least a hundred distinct years and still appearing in the latest release, ordered by lifetime births. This is the closest the data comes to timeless.

Comeback: names whose peak sits at least half a century behind, that are still in the data, and whose ten-year trend exceeds plus ten percent. This is the reuse cycle caught in the act.

Unisex balance: for each spelling that appears in both columns with at least 5,000 births each, we compute the smaller column divided by the larger — 100 percent would mean identical usage. We rank by combined usage so the chart shows balanced names that people actually choose.

What we do not do: we do not weight, smooth, impute suppressed values, blend other datasets, or rank quality of any kind. Where the source is silent — the five-baby floor — our charts are silent too. When the SSA publishes each May, this site's totals recompute wholesale; the About page stamps the data vintage.