Sources and method

Yıldaşlar is a statistical photo of 1,000 people born the same year as you. Dots are not real people; each dot stands for one in a thousand of those born that year.

Every number comes from quotas computed from the open sources below. The same year and country always produce the same photo; the shared card is exactly what you saw on screen.

Method

1. Name era and guessing style. A name's era is the set of decades in which it reached its best observed rank in TurkStat's 1950–2025 top 100 lists; gaps between peaks are not filled. Names peaking in at least three decades are labelled 'Across generations'. This is not the average age of people with that name; older lists include decade estimates. Three questions use the package's first-year losses, birth shares by continent and survivors today in reference countries. All correct earns 'Statistics hunter'; other completed rounds earn 'Curious guesser'. With a drawn curve, the average future difference uses a ±15-person band to assign optimistic, cautious or close to the projection. The hunter label takes priority; the curve's direction is also shown. A shared link includes the name, name-list preference, answers and any curve values, which are recalculated for the same class and data version on the recipient's device.
2. Birth counts. Türkiye's 2001–2025 birth counts are TurkStat's recorded annual totals, summed from the monthly birth table; earlier years and other countries use UN WPP 2024. Survival still uses UN life tables; correcting birth counts does not change mortality rates.
3. Birthplaces in the world class. Annual births for 237 countries/areas in UN WPP 2024 are scaled by the largest-remainder method to country shares summing to exactly 1,000; continent counts are sums of those shares. North and Latin America form one group; Türkiye is in Asia under the UN classification. All shares, including Türkiye's, use WPP in this scene for consistency across countries. The six largest countries are listed. Classes at least 15 years before 2025 compare with 2025; more recent classes compare with 1950.
4. Generation markers. The entrance-exam name is based on the exam year when the class turned 18 (ÖSYM). Since birth months are unknown, the first ballot is the first nationwide vote at which the entire class met the age rule; survivors are counted at the start of that year. The 1987–1995 rule included those turning 20 in the election year; after 1995, voters had to have turned 18 on election day. Silent–Gen Z boundaries follow Pew; Alpha 2013–2024 is an editorial range used here, and Beta 2025–2039 follows McCrindle. These are not scientific groups specific to Türkiye.
5. Draw your own curve. The past uses the class's same survival array. You draw from today onwards; gaps are interpolated linearly and the undrawn tail falls to zero by the end of the final age year. The curve cannot rise or exceed today's survivor count. If more than half the class is alive today, the age when half remains is compared with the UN projection; otherwise survivors at a milestone about 15 years ahead are compared. The optimistic/pessimistic line uses the average future difference (within ±15 people is considered close). The drawing stays in this page session; if you choose to share a result link, its curve values are included in that link.
6. Cohort survival. The UN's single-age period life tables for 1950–2100 are chained diagonally: for birth year Y and age x we use the mortality of year Y+x. Rates are held constant after 2100.
7. Beyond age 100. The tables end at 100; ages 101–110 are extended with a logistic (Kannisto) curve fitted to ages 80–99 and closed at 110. That is why the 'last of you' is given as a ±5-year range rather than a single year.
8. Exact counts, not random draws. The expected number of survivors at each age is cumulatively rounded and deaths are assigned to exactly that many dots. The dots always sum to 1,000 and the picture matches the statistics exactly.
9. Quotas stratified by sex. The sex ratio at birth sets the number of boys and girls; death, education and marriage quotas are computed within each sex and among those alive at that age. Which dot is chosen comes from a seed derived from country + year + data version.
10. Education and marriage hierarchy. Post-secondary graduates are a subset of upper-secondary graduates, who are a subset of those who started school; marrying dots are drawn from those alive. About seventy percent of marriages before 18 are drawn from women with no or primary schooling (an editorial assumption).
11. You. The red dot is chosen among the dots alive today with a seed private to your browser; in the future scene it leaves the quota pool and never turns grey. About you, the text states only what has already happened ('you made it to today'); the future always belongs to the class.
12. Comparisons. 'Your mother's class' is 27 years before your birth year and 'your child's class' 27 years after (the mean age of mothers in Türkiye is 27–29; we do not know the actual year). 'Newest class' is the last birth year in the data package. The world line is the UN world total, not a country-by-country comparison. The size rank orders the classes in the package by number of births.
13. Country lottery. The class of the same year is built with the same method for eight reference countries (Japan, Germany, United States, Russia, China, India, Nigeria, Brazil; UN life tables, exact counts by sex). The table shows the class with the most survivors today and the one with the fewest; ties go to the class with fewer deaths in year one. The eight were chosen to sample the world's largest populations and most different survival paths; the ranking is not a judgement but a measure of what the same birth year means elsewhere.
14. How much longer? The optional line is the class's average remaining lifetime from its current age: the sum of survivors at all later ages divided by the number alive today, plus 0.5 for deaths within the year (e_x = Σ S(a>x) / S(x) + 0.5). It is computed separately for women and men from the same death arrays. It is a class average, not a personal forecast; it never opens by itself and is not offered for classes younger than 18.
15. Doors. 'Another life' opens three doors: another year (another class of the same country), another country (the same year's class in one of eight reference countries or the world, built from the UN life tables with the same method) and the other sex (the same class's women and men scaled to 1,000). The comparison is always made at your class's current age; if the other class has not reached it, the figures say 'projection'. 'The dot in your place' is drawn from the other class with a seed specific to your browser and need not be alive today; behind the door you may have died. Door data is downloaded only when a door is opened; the 'I'd stay / I'd take it' answer is counted only in your browser and never sent to a server.
16. Fork in the road. 'Fork in the road' builds a counterfactual past from three choices: type of high school (none / general / vocational / imam hatip), marriage and work. How many went to high school is the cohort's own education share; the split into types uses the Ministry of Education's current student shares (estimate). Going on to higher education is a lottery: ÖSYM placement rates by school type are taken as relative weights and shifted on the odds scale to match the cohort's own tertiary share, so the class totals do not change. At the marriage fork the dots' marriage ages are real data; at the work fork labour-force participation by education (TurkStat) and occupation-group shares are used. The card speaks up to today's age: age at first marriage is drawn from the distribution by education truncated at today's age, divorce is cumulative by marriage duration, children scale total fertility by mother's education to the years married, earnings are the occupation group's 2023 annual gross average given as a multiple of the minimum wage. The only forward-looking line is remaining life at 30 by education (TurkStat life tables), a class average. Lotteries are seeded by the class and the 'you' dot: a shared link shows the same card. 'I'd stay / I'd take it' is counted only in your browser.
17. The yearbook. Notes at the end of a class page are kept separately for each year and country and are anonymous: no name, account or email is asked for. Your browser creates a random identifier and keeps it only on your device; the server derives a four-character tag (e.g. #K3F2) from a salted hash of it and lets only that device delete the note. IP addresses are not stored; to limit abuse, a salted hash is kept for at most 24 hours in posting counters and for 90 days with reports. Notes containing links, phone numbers, emails, usernames, swearing or insults are not published; a note reported by three different people is hidden. The newest 500 notes are kept per class.
18. Life events. Every year of the flow has an event of the year: deaths and marriages are counted from the class's own quotas; first child, first car and first home are a model and carry the 'est.' label. First child: every dot that marries gets a first child 1–4 years after the wedding (a distribution calibrated to TurkStat's mean age of mother at first birth minus women's mean age at first marriage, ≈1.4 years); 6% of the married stay childless (8% in TDHS 2018, lowered for completed fertility), and women's cumulative first children never exceed the class's total children in the UN fertility data. First car: cars per 1,000 inhabitants by year (TurkStat/Eurostat, 1950–2025) times average household size, less second cars, gives the share of households with a car; it is multiplied by an age profile starting at 18 (peak at 35–54). First home: an eventual home-ownership level by birth decade (from Eurostat EU-SILC age groups in 2015 and today) times the age distribution of first-time buyers (RE/MAX 2025: 20% at 18–25, 36% at 26–35; the breakdown after 35 is an editorial completion, up to 3 years earlier for older cohorts). Each dot gets a single seeded threshold and lives the event at the first age where the model's cumulative probability passes it, while alive; so each year's counts follow the model curve exactly and the same class always gives the same years. The 'majority' line is the first year in which more than half of those alive had lived the event. Türkiye has no official age distribution for first cars or first homes; read those two lines as a sourced estimate, not a measurement. 'Do you have one?' answers stay in your browser only.
19. The 'estimated' label. Lines based on model estimates, decade averages or projections are marked 'est.': mortality 1950–1989, names 1950–1979, birth months before 2001, marriages after 2013, education after 1996, and all first-child / first-car / first-home lines.
20. Workdays until 65. The count runs from today to the 65th birthday; without a birthday, mid-year (1 July) is used. A workday is a weekday; for the Türkiye class, full-day public holidays are subtracted (1 January, 23 April, 1 May, 19 May, 15 July, 30 August, 29 October, three days of Ramazan Bayramı and four of Kurban Bayramı). Religious holidays come from the tabular Islamic calendar: it matches 18 of the 24 holiday dates in 2015–2026 exactly and the rest within one day. Half days and annual leave are not counted; no holidays are subtracted for the world class. The share of working life behind you assumes a start at age 22 unless you enter a year, counted from mid-year. 65 is a reference age, not a retirement age. Those who reach 65 are the same dots as in the class photo; where 65 lies in the future it is the UN projection. Your birthday and start year are stored only on your device and never sent to the server.

Data sources

Data package version 2026.11.3; manifest generated on 9 November 2026.

Limits and known gaps

• Dots are a quota sample, not individuals; two people from the same class see the same picture.
• TurkStat's new Data Portal has no education or marital-status tables by age, so education relies on the Wittgenstein Centre and marriage on the Turkey Demographic and Health Surveys.
• The men's marriage curve is derived from the women's curve using the gap in mean age at first marriage.
• Names are the ranking of those alive today in the population register; decade rows are used for the 1950–1979 cohorts.
• Birth-month distribution for cohorts before 2001 is the 2001–2010 average; zodiac shares assume an even spread within each month.
• Migration, prison and causes of death are not in this version. The World class uses the UN's world-total life table.
• In the fork in the road, school-type shares are a 2025-26 snapshot applied to every cohort; ÖSYM rates are single-year; earnings are 2023, marriage/divorce/fertility 2024, labour force and happiness 2025. On a man's road the number of children assumes the spouse's education equals his own; the happiness share is not conditioned on age.

How it was made

The data pipeline is written in Python (pandas, pyarrow); every run writes source checksums and retrieval dates to this page. The interface is Next.js and TypeScript; the 1,000 dots are drawn with Canvas 2D; share images come from the same engine.

Typeface: Inter (SIL Open Font License). Cookie-free analytics. Yearbook notes are anonymous; no names, emails or IP addresses are stored.