Births and marriages by wilaya — Algeria (2013-2019)

Live births and marriages in Algeria, quarter by quarter and wilaya by wilaya, 2013 to 2019. 1,056 observations, 22 quarters, all 48 wilayas present in each, births broken down by sex.

What you cannot do without this file. Algerian vital statistics are released annually at national level. This granularity — quarterly AND by wilaya — exists only in the Quarterly Statistical Bulletins, where it occupies one page out of more than a hundred, between rainfall and price indices, in a bilingual document whose text layer interleaves French and Arabic. Twenty-eight publications were needed to rebuild these twenty-two quarters.

Each bulletin carries two quarters, which makes cross-checking possible. The table gives the current quarter and the same quarter a year earlier. The second quarter of 2018 is therefore published once by the 2018 bulletin and again by the 2019 one as a comparison column — two documents, two pages, two readings. Across the 432 observations seen twice: 432 agreements, no divergence. The confirme_par_deux_bulletins column marks them.

An arithmetic check on every row as well: male plus female equals total births, wilaya by wilaya. And a completeness check: 48 wilayas expected, 48 found in each of the twenty-two quarters — a table returning 47 would have lost a row somewhere, and nothing else would say so.

What is not published, and why. Marriages for two quarters — the fourth of 2015 and the second of 2016 — are absent. The same cross-check that validates births finds these columns wanting: 95 divergences, all concentrated in those two quarters, with the ratio between the two readings running from 0.01 to 25. A vital-statistics revision does not behave that way; this is a column we failed to align. Since we do not know which reading is right, neither is shipped. Publishing one at random would have been worse: nothing in the file would have flagged the doubt.

Three reading traps, mentioned because they say what the file avoided. Some bulletins publish their current quarter as ellipses and give figures only for the previous year — reading them naively dated 2018 births as 2019, while perfectly satisfying the arithmetic check. The fourth-quarter bulletin carries two identically titled tables, the quarter and the annual cumulative — confusing them published 962,485 births where a quarter holds 250,000. And the dash in "Tizi – Ouzou" comes through mis-encoded in recent editions, making wilaya 15 disappear.

Coverage. 2013-Q1 to 2013-Q3, all of 2014 and 2015, all of 2016 and 2017, then 2018-Q2, 2018-Q4 and 2019-Q2. Missing quarters correspond to bulletins not published in this form, or whose table does not extract cleanly — they are not filled in by interpolation.

Seasonality is clear and consistent year to year: the third quarter is always the most fertile, at 275,000 to 300,000 births against 224,000 to 260,000 for the others.