In 1977, 1,117 American girls were named Kizzy.
The year before, the name did not appear in the national baby-name data at all.
Then Roots aired.
Kizzy was the daughter played by Leslie Uggams in Roots, one of the largest television events in American history. The name did not climb. It arrived.
That is the idea behind Debut of the Year: for every year, find the boy name and girl name with the largest count in their first appearance on the national list.
Open the interactive Debut of the Year almanac →
The result is not a ranking of the best new names. It is a record of sudden transmission. A name was absent, or too rare for the Social Security Administration to report. Then something happened. A battle entered the newspapers. A television character entered millions of living rooms. A singer found an audience. A migration brought a name into a new national record. A spelling escaped from one family and became a pattern.
Sometimes the cause is obvious.
Often it is not.
The screen made names move at national speed
The early winners are small. In 1916, Verdun debuted with 14 boys during the battle that consumed France. In 1918, Foch arrived with 58, after Marshal Ferdinand Foch became the face of Allied victory. In 1931, Rockne appeared after the death of Notre Dame coach Knute Rockne.
These names came through headlines.
By the 1950s, the screen had taken over. Maverick debuted when the television western premiered. Rowdy followed Rowdy Yates, Clint Eastwood's character on Rawhide. Illya arrived with the beautiful Russian spy from The Man from U.N.C.L.E.
Then television demonstrated its full naming power.
The ABC police drama Nakia premiered in 1974. That year, Nakia became the largest valid boy-name debut in the almanac: 613 births, distributed across major television markets.
Three years later came Roots. Levar debuted with 523 boys. Kizzy debuted with 1,117 girls, nearly twice Nakia's count.
The scale matters. Earlier headlines could introduce a name. Network television could synchronize it.
Girls' names made the project better
The almanac began with boys.
That version already had a clean story: war heroes became television cowboys; television cowboys gave way to athletes, musicians, rappers, reality stars, and Spanish-language media figures.
Then we ran the same query for girls.
The girls' list did not repeat the story. It widened it.
There was Monalisa in 1950, the year Nat King Cole's "Mona Lisa" ruled the charts. Phaedra followed the Sophia Loren film. Fallon arrived with Dynasty. Sade appeared at scale during the singer's American breakthrough. Moesha, Erykah, Kehlani, and Zhavia each turned a newly prominent woman or character into a birth-certificate event.
The Spanish-language thread became clearer too. Isamar debuted with 447 girls in 1990. Aideliz and Greidys arrived in the years those women appeared on Nuestra Belleza Latina. Dalary spread after singer Larry Hernández introduced his daughter publicly.
The girls were not an appendix. They were half the evidence.
A debut is not the same thing as invention
The word debut needs care.
The Social Security Administration suppresses names with fewer than five births in a year. A name's first appearance in the dataset therefore means its first year at or above the reporting threshold, not necessarily the first time any American parent used it.
Some winners are old names crossing that threshold. Some are surnames moving into first-name use. Some are phonetic spellings of familiar names. Some were already established inside a community before the national data could see them.
And some appear to be mistakes.
In 1928, the winning boy debut was Alfread. In 1929, it was Donnald. In 1938, Daivd.
Those spellings remain because they are present in the underlying SSA record. A typo on a birth certificate is still part of the naming record.
The 1989 winners are different. Christop appears for 1,082 boys and Alexandr for 301 girls. Both are eight-character truncations, artifacts introduced by a source-data constraint rather than names reported as written. They sit in red in the almanac: not winners to celebrate, but evidence that data can carry the shape of the machine that handled it.
Geography tells you where, not why
The hardest entries are the ones without an obvious headline.
The database has an advantage that ordinary name lists do not: debut-year counts by state. Those counts produce geographic fingerprints.
Mychal Thompson's 1978 debut is led by Minnesota, where he played college basketball before becoming the first pick in the NBA draft. Spanish-language media debuts often lean toward California, Texas, and Arizona. Some names begin in the Deep South or the urban Northeast. Others appear almost evenly across the country, the footprint one might expect from national television.
This evidence is useful. It is also dangerous.
A California-Texas cluster can suggest a Spanish-language cultural channel. It cannot identify a telenovela character by itself. A broad national spread can be consistent with network television. It does not prove that a television show caused the name. Geography can narrow the search. It cannot complete it.
The distinction became one of the central editorial rules of the project:
Geography is evidence of where a name caught on, not proof of why.
That sentence exists because the first drafts repeatedly tried to make geography do more than it could.
How the almanac was made
This project began the way many data stories begin now: one person with a database, several language models, and a lot of back-and-forth.
I had the underlying name data. An early query produced the largest boy-name debut for each year. One model looked at the list and saw the shape immediately: war heroes, television westerns, civil-rights figures, hip-hop, reality television, telenovelas. "A century of American media history written in birth certificates" became the working idea.
The first instinct was to annotate every year.
That was the wrong instinct.
Obvious entries were easy. Foch, Maverick, Medgar, Nakia, Levar, Pharrell, Jionni, Vanellope. The danger began in the blank spaces. Language models dislike blank spaces. Give one a name, a year, and a request for an explanation, and it will usually find a story-shaped object to put there.
DeepSeek's pass became valuable precisely because its normal web-search route failed. It stopped searching the open web and inspected the database itself. Its own debrief called that pivot the decisive move.
It read the schema first. That exposed the state-level table, the regional-anomaly table, the known-catalyst table, and the main name records. It compared debut-year state distributions. It searched spelling neighborhoods instead of treating every spelling as an isolated object. It looked at the full cohort of names arriving in the same year. Most importantly, it eventually recognized when the database had nothing more to say.
Those moves changed the work.
Shelva is the best example. The name did not appear alone. Shelba, Shelvia, Shelvie, and several neighboring spellings rose together in 1936 and 1937, heavily concentrated in Appalachia. That pattern suggests a name transmitted by sound: people heard something similar and wrote it several ways.
But the model went one step farther. It declared the source an Appalachian radio character.
There may have been one. We did not find one.
The data supported oral transmission and regional concentration. It did not support the radio character. The published note stops where the evidence stops.
That correction came from a second kind of model work. Codex approached the annotations as a skeptical editor rather than an attribution engine. It separated observation from inference, checked confident claims against stronger alternatives, and repeatedly asked whether geography established a cause or merely suggested where to look.
That back-and-forth was useful. DeepSeek widened the field: new tables, spelling neighborhoods, geographic patterns, cohort comparisons, possible catalysts. Codex narrowed it again: unsupported radio characters came out, generic "TikTok" explanations came out, and several vague telenovela guesses were replaced with identifiable people or with honest uncertainty.
That became the real collaboration. The models generated hypotheses, challenged one another's conclusions, found database paths I had not considered, and accelerated the search. I decided which claims had earned a sentence.
The useful disagreement
Different models saw different things in the same evidence, and they brought different temperaments to it.
One would classify a name as part of a broad cultural naming pattern. Another would hunt for a specific actor. Another would notice that a neighboring spelling debuted in the same year. Sometimes one model's confident attribution gave the next model a target to disprove.
The disagreement was worth it.
For Djuna, a database-only interpretation placed the name inside a larger wave of new D- and De- constructions. Further research found Djuna Phrayne, a character on the television drama Channing during the 1963–64 season. The simultaneous debut of Dejuna still matters: the name may have entered through television and then been reshaped as people heard it.
For Greidys, geography suggested Spanish-language media. That was directionally right but not specific. The actual catalyst was stronger: Greidys Gil won Nuestra Belleza Latina in 2009, the same year the name debuted with 187 girls.
That was where the models worked best together. One found a pattern. Another found a person. The database checked the timing and scale. Editorial judgment decided how firmly to connect them.
The mysteries are part of the result
The finished almanac does not explain every name.
That is not a defect.
Some early debuts are too small and too old to leave much evidence. Some newer names spread through social platforms, family networks, and communities whose traces are difficult to search. Some names may have had a local catalyst that never entered a durable archive. Some may simply be inventions that occurred independently in several families.
The temptation is to turn every correlation into a cause. The better story includes the limit.
A name can have a year, a count, a spelling neighborhood, and a geographic footprint, yet still resist explanation.
The record tells us that it happened.
It does not always tell us why.
What the winners show
Read across the full century and the pattern is not that media started influencing names. Headlines, songs, performers, and public figures were present from the beginning.
What changed was the machinery of attention.
Newspapers could make Foch visible. Radio could make a name audible. Network television could deliver Nakia or Kizzy to tens of millions of households at once. Cable, hip-hop, reality television, Spanish-language broadcasting, and social media fractured that audience into many overlapping publics.
The winners become more culturally varied because American attention became more culturally varied.
The almanac is also a history of how Americans came to know the same person, character, sound, or spelling at the same time.
Sometimes the country watched together.
Sometimes a community moved first.
Sometimes a name arrived and left no explanation behind.