Birthdays
How common is your birthday?
Every date of the year, ranked by how many people were born on it.
Point at a square to read it.
Each square is one date, shaded by how far its births sit from an average day of the same year. The middle step is a date within 2% of an average day. The darkest step on the quiet side holds every date below −15.0% — 6 of them, down to February 29 at −77.7%, which is far off the scale.
What the counts say
9 of the ten most common dates fall in September, led by September 9 with 12,301 births a year — +10.6% against an average day. September is the busiest month of the year: its dates average +5.7%.
The bottom of the table is holidays. December 25 (6,574 a year, −40.8%), January 1 (7,792 a year, −29.9%), December 24 (8,069 a year, −27.4%) and July 4 (−20.9%) are the four quietest dates that exist every year. Below them all sits February 29, which existed in only 5 of these 21 years: 2,491 a year averaged over the whole period, and −6.2% when it does come round.
The week moves the count as much as the calendar does. Tuesday averages +14.8% against an average day and Sunday −31.5% — a gap of 46 points, against 51 points between the busiest and the quietest date of the year. Weekdays run +10.9%, weekends −27.1%.
The ten most common dates
The ten rarest dates
The week beats the calendar
Every one of the 7,670 days in the data, grouped by the weekday it fell on. The figure is that day’s births against the average day of its own year.
| Weekday | Births on such a day | Against an average day | |
|---|---|---|---|
| Monday | 11,563 | +3.8% | |
| Tuesday | 12,788 | +14.8% | |
| Wednesday | 12,569 | +12.8% | |
| Thursday | 12,503 | +12.3% | |
| Friday | 12,304 | +10.5% | |
| Saturday | 8,593 | −22.7% | |
| Sunday | 7,610 | −31.5% |
Two people in a room
Arithmetic, not a finding in the data: the chance that at least two people in a group share a birthday, worked out from the shares above. The second column is the textbook figure, which assumes all 365 dates are equally likely.
| People in the room | With these dates | If dates were even | |
|---|---|---|---|
| 5 | 2.7% | 2.7% | |
| 10 | 11.7% | 11.7% | |
| 20 | 41.2% | 41.1% | |
| 23 | 50.8% | 50.7% | |
| 30 | 70.7% | 70.6% | |
| 40 | 89.2% | 89.1% | |
| 50 | 97.1% | 97.0% | |
| 70 | 99.9% | 99.9% |
It takes 23 people for the chance to pass a half: 50.8% with the real spread of dates, 50.7% with even ones. Dates are not quite even, which nudges the chance up, never down.
About the data
What exactly is being counted?
Registered live births in the United States, by the exact calendar date they happened, for the 21 years from 1994 to 2014. 85,386,227 of them. It is a count of births that already happened — nothing on this page is advice, a prediction or a medical statement of any kind.
Why does the data stop in 2014?
Because that is where the public files with day-of-month detail stop. The CDC’s own query tool, WONDER Natality, covers later years but lets you group births only by year, by month and by weekday-or-weekend — not by day of the month. Rather than extend a series past its data, the section says 1994–2014 and stops there.
Why is 29 February last?
Because it existed in only 5 of these 21 years. Averaged over the whole period it comes to 2,491 births a year, far below any date that happens annually. In a leap year itself it is an ordinary date, running −6.2% against the average day of that year.
How is the rank worked out?
For each of the 21 years, a date’s births are divided by the average day of that same year; the 21 figures are averaged, and the dates are sorted by the result. Dividing by each year’s own average takes out the growth of the population and makes the two source files comparable. A date that does not exist in a year counts as zero for that year, which is why 29 February finishes last.
Is the weekend dip real?
It is in the counts, and it is large: the five weekdays average +10.9% against an average day, Saturday and Sunday −27.1%. This site reports the counts and does not explain them.
Can I use these numbers?
Yes. The underlying federal tabulations are in the public domain; the copies used here come from FiveThirtyEight’s data repository under CC BY 4.0, so credit FiveThirtyEight as well as the CDC/NCHS and the SSA. The exact files, with their checksums, are named in the script that builds this section.
Where the numbers come from. Births by exact date, 1994–2014: 1994–2003 from the Centers for Disease Control and Prevention, National Center for Health Statistics (National Vital Statistics System); 2004–2014 from the Social Security Administration. Both files are works of the US government and in the public domain. They are taken from FiveThirtyEight's data repository, which republishes them under CC BY 4.0 — credit to FiveThirtyEight, github.com/fivethirtyeight/data.
The two files overlap in 2000–2003, where the SSA count runs about 2% above the CDC/NCHS one; those four years are taken from the vital-statistics file, so each year has one source. Every date is scored against the average day of its own year before the years are averaged, so the join cannot move a ranking.