The archive
Every number this site publishes comes from here. 54,901 rows of Philippine hog and pork prices, herd size and disease data going back to January 1945, and all of it is free to download.
Most pages you will find on presyo ng baboy quote one figure from one press release and do not say which week it was. That is not really their fault — the underlying numbers are scattered across a PXWeb query interface and 109 separate PDFs, and nobody had put them in one place. So we did, and then we published the working.
01 · In numbers
Every figure on this page is counted from the archive during the build. None of them is typed, which is the only way a claim like this stays true after the next refresh.
Rows
54,901
Live rows across every series and every geography.
Carrying a figure
40,314
73% of them. The rest are dated absences.
Published absences
14,587
Periods the source printed and left blank. Kept, dated, never filled in.
Withdrawn
30
Rows this site published and could not stand behind. Marked, never deleted.
Series
59
Each with its own page and four downloads.
Geographies
112
18 regions and 92 provinces, plus the nation.
Documents retrieved
173
Each hashed on the exact bytes received, and re-findable by that hash.
Years spanned
81
1945-01-31 to 2026-07-31.
02 · The four rules
RULE 1
Nothing here is interpolated, averaged into place or carried forward from last time. 14,587 of these rows carry no figure at all, because the source published none — and each of those rows says which source, which period, and when somebody looked. A gap you can see beats a smooth line you cannot check.
RULE 2
Not the publisher, the page — with the URL and the date it was actually opened. A figure whose provenance is "PSA" is a figure nobody can go and check.
RULE 3
PSA and BAI both survey hog farmgate prices and they do not match. Averaging them would produce a price no survey measured, so both are here as separate series and the difference is published as its own.
RULE 4
When we get something wrong, the row is marked withdrawn and stays, with the date and the reason. 30 rows in this archive are marked that way, most of them from a July 2026 audit that found figures on this site with no source behind them at all. That record is on the series pages, and it is there on purpose.
03 · Start here
Each of these carries a figure rather than a promise, because a tile that only says where it goes is navigation and this page is meant to be evidence.
What a pig fetches where. The dearest region and the cheapest are far enough apart that a national average is not a planning number.
18 regions and 92 provinces, each with its own line
PSA surveys both and subtracts neither. The gap between them, month by month since 2010, with the arithmetic shown.
150 months differenced, 2010-01-31 to 2026-07-31
Every document read to build this, with the SHA-256 of the exact bytes. Download one and check the hash yourself — it is the one claim here you can disprove without trusting us.
173 documents retrieved, each content-addressed
Every frozen release with the correction it makes against the last, and every retrieval that brought in a document this archive had never read. There is an Atom feed.
5 frozen releases, and an Atom feed of the retrievals
The numbers this archive cannot give you, what was checked before saying so, and what we publish instead. A gap with no search behind it is an assumption.
10 open records, 6 of them silencing a series here
The URL shapes, the column dictionary and worked snippets for pandas, DuckDB and R. No key, no rate limit, no endpoint — these are files.
4 formats per series, 237 files in all
04 · The catalogue
Filter with the chips. It is CSS — no JavaScript runs to hide a card, the counts are computed during the build, and every series stays in the HTML whatever is selected.
10 series · 982 rows
The price at the farm gate, per kilo liveweight. Two agencies survey it and they disagree, so both are here as separate series and the difference between them is published as its own.
172.44 PHP/kg liveweight · Q2 2026
-6.8%
The curated quarterly figure this site quotes as its headline price, assembled by hand from PSA releases and trade reporting before the machine-retrieved series below existed. It also carries this archive's retraction record: rows the site once published, could not source, and withdrew in the July 2026 audit.
Philippine Statistics Authority — Swine Situation Report
171.76 PHP/kg liveweight · 2026-06
+0.1%
PSA's monthly farmgate price for hogs for slaughter across all farm types. The newest of PSA's three populations — this table begins in 2025, where the other two reach back to 2010.
PSA OpenSTAT — Philippine Statistics Authority statistical database
172.62 PHP/kg liveweight · 2026-06
+0.5%
PSA's monthly farmgate survey of household farms — the backyard population most people reading this site belong to. Monthly since January 2010, and the longest continuous price series here.
PSA OpenSTAT — Philippine Statistics Authority statistical database
170.52 PHP/kg liveweight · 2026-06
-0.4%
PSA's monthly farmgate survey of commercial establishments. The same commodity as the backyard series and a different population: the two are never averaged, and the gap between them is published as its own series instead.
PSA OpenSTAT — Philippine Statistics Authority statistical database
199.50 PHP/kg liveweight · as of 2024-07-22
-8.1%
BAI's own weekly farmgate price for backyard hogs — a second agency measuring what PSA measures, at four times the cadence. BAI stopped printing a figure in this column after 22 July 2024 and has printed N/A every week since; the dated blanks are kept, because when a source went quiet is itself a fact.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
227.50 PHP/kg liveweight · as of 2025-08-25
0.0%
BAI's own weekly farmgate price for commercial hogs. It ran thirteen months longer than the backyard column beside it and stopped the same way, after 25 August 2025.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
172.44 PHP/kg liveweight · Q2 2026
-2.0%
The quarterly mean of PSA's all-farms monthly rows, computed from this archive with the arithmetic attached to every row. A quarter missing a month has no mean and says which month is missing.
Baboy PH editorial synthesis
172.21 PHP/kg liveweight · Q2 2026
-3.9%
The quarterly mean of PSA's backyard monthly rows. This is the shape the headline figure is quoted in, and computing it here rather than by hand is what makes a stale headline something a gate can notice.
Baboy PH editorial synthesis
172.77 PHP/kg liveweight · Q2 2026
+0.8%
The quarterly mean of PSA's commercial monthly rows. Sixteen quarters between 2010 and 2013 have no mean at all, because PSA published nothing for those months and the average of the months it did publish is a different quantity.
Baboy PH editorial synthesis
2.10 PHP/kg liveweight · 2026-06
+296.2%
Backyard farmgate minus commercial farmgate, monthly, computed from two PSA populations that neither agency subtracts from each other. A difference between two sources is a real artifact; an average of them would be a price nobody surveyed.
Baboy PH editorial synthesis
3 series · 45,870 rows
The same surveys across every region and province PSA reports. This is where a national average stops being a planning number.
224.10 PHP/kg liveweight · 2026-06
+6.2%
The all-farms survey across every Philippine region and province. Every row carries its own PSA geolocation code, because a provincial figure that reaches a national chart unlabelled is undetectable in review — it just looks like a month where the price moved.
PSA OpenSTAT — Philippine Statistics Authority statistical database
224.20 PHP/kg liveweight · 2026-06
+6.2%
The backyard survey across every Philippine region and province, monthly since 2010. The largest thing in this archive, and the reason a single national average is close to useless as a planning number: the dearest region and the cheapest are further apart than most people think the whole market moves in a year.
PSA OpenSTAT — Philippine Statistics Authority statistical database
200.54 PHP/kg liveweight · 2026-06
-0.6%
The commercial survey across every Philippine region and province. Close to half its rows carry no figure, and that is not a gap in the retrieval: PSA publishes a provincial commercial price only where there are establishments to survey, and the blanks are where there are not.
PSA OpenSTAT — Philippine Statistics Authority statistical database
9 series · 952 rows
BAI and PSA both read these out of weekly and monthly returns. Retail liempo runs roughly twice the farmgate price of the same week, which is what most arguments about pork prices are actually about.
332.07 PHP/kg, PSA retail pork kasim · 2026-07
-1.3%
PSA's monthly national average retail price for fresh pork kasim. Set against `hog-farmgate-average` it gives the farmer's share directly, and the gap is wider than most people assume.
PSA OpenSTAT — Philippine Statistics Authority statistical database
349.10 PHP/kg, PSA retail pork liempo · 2026-07
-1.3%
Monthly national average for fresh liempo, the dearest of the four cuts PSA prices and the one that moves most with import volume.
PSA OpenSTAT — Philippine Statistics Authority statistical database
302.43 PHP/kg, PSA retail pork meat with bones · 2026-07
-1.4%
Monthly national average for fresh pork sold bone-in. The cheapest of the four, and the one most households actually put in the basket.
PSA OpenSTAT — Philippine Statistics Authority statistical database
278.27 PHP/kg, PSA retail pork pata (fore) · 2026-07
-1.1%
Monthly national average for fresh front pata. Worth keeping because it tracks the other cuts least closely, which makes it a check on whether a swing is the whole market or just one cut.
PSA OpenSTAT — Philippine Statistics Authority statistical database
260.00 PHP/kg, wholesale sabit ulo · as of 2026-07-13
0.0%
BAI's weekly wholesale price for sabit ulo, read out of the price monitoring PDF one dated report at a time. A wholesale price, several market levels above the farmgate price a raiser is paid — never the same line.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
378.00 PHP/kg, retail pork liempo · as of 2026-07-13
-0.5%
What liempo costs at retail, weekly, from BAI's price monitoring. This is the number that shows up in the news when pork prices are a story, and it is roughly twice what the farmgate series beside it says a raiser was paid the same week.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
344.00 PHP/kg, retail pork kasim · as of 2026-07-13
+1.2%
What kasim costs at retail, weekly, from BAI's price monitoring. Cheaper than liempo every week of the archive, and the cut most households actually buy.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
300.00 PHP/kg, retail frozen liempo · as of 2026-07-13
0.0%
BAI's weekly price for frozen liempo, which it prices in its own table beside the fresh cut. A different product at a different price, not a second opinion on the same one.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
240.00 PHP/kg, retail frozen kasim · as of 2026-07-13
-4.0%
BAI's weekly price for frozen kasim. The cheapest pork in the report most weeks, and the one imports compete on.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
8 series · 301 rows
PSA counts the national herd quarterly, and changed how it splits that count after 2020. Both classifications are here and nothing joins them.
8.93 million heads · 2026
+1.6%
PSA's annual head count of the national herd, all classes. One reading a year, and the slowest-moving number on this site.
Philippine Statistics Authority — Swine Situation Report
8,928,329 heads · 01 Jul 2026
+2.6%
PSA's count of every pig in the country on the first morning of each quarter, all farm classifications together, from 01 January 2010. It measures the herd rather than what the herd costs, which nothing else here does.
PSA OpenSTAT — Philippine Statistics Authority statistical database
7,080,312 heads · 01 Jul 2026
+3.6%
The backyard herd counted on its own. Close to four fifths of the national inventory sits in this classification, which is why a commercial-farm figure quoted as a national one is wrong by a margin nobody checks.
PSA OpenSTAT — Philippine Statistics Authority statistical database
149,216 heads · 01 Jul 2026
+43.4%
The smallest of the three classifications and the one that swings hardest. PSA created it when it retired the old Backyard/Commercial split, so it has no counterpart in the 2013-2020 series.
PSA OpenSTAT — Philippine Statistics Authority statistical database
1,698,801 heads · 01 Jul 2026
-3.7%
The commercial herd on the same quarterly morning. Read beside the smallhold count rather than instead of it: the two populations came back from African swine fever at visibly different speeds.
PSA OpenSTAT — Philippine Statistics Authority statistical database
12,795,721 heads · 2020
+0.7%
The national herd as counted on the first morning of January each year, under the farm-type classification PSA used until 2020. Its own highest reading is a January 2020 one, after ASF arrived and before it reached the commercial sector in force. It is NOT the highest the national herd has been: the successor series begins three years earlier and counts higher, which is why nothing here calls this table a peak. PSA still labels its final year preliminary and has not published the table since; the successor series counts by farm size instead and the two schemes must not be joined.
PSA OpenSTAT — Philippine Statistics Authority statistical database
7,971,400 heads · 2020
-2.4%
The backyard half of the retired classification. Not the same population as the smallhold series, which is exactly why they are kept apart: PSA moved the boundary when it replaced the scheme, so a farm that counted as backyard in 2020 is not necessarily smallhold now.
PSA OpenSTAT — Philippine Statistics Authority statistical database
4,824,321 heads · 2020
+6.2%
The commercial half of the retired classification, 2013 to 2020. Its successor splits the same farms across Semi-Commercial and Commercial, so the old and new lines do not meet and are never drawn as one.
PSA OpenSTAT — Philippine Statistics Authority statistical database
9 series · 2,580 rows
Production and slaughter from PSA, and the import, export and stock side of the balance sheet from USDA. Different publishers, different definitions, never added together.
412,282 metric ton, liveweight · 2026-Q2
-3.9%
Quarterly hog production in liveweight tonnes, back to 1997. The commodity sits two dots deep in PSA's dimension: `LIVESTOCK1/` at depth zero is the sum over hogs, cattle, carabao, goats and dairy, and reading that as hogs overstates production by roughly a third.
PSA OpenSTAT — Philippine Statistics Authority statistical database
5,455 metric ton · 2026-Q2
+18.4%
The same quarterly volume across all eighteen regions, back to 2000. PSA declares its unit here as plain metric ton where the national table of the same commodity says metric ton liveweight, so the two are charted together only with that difference stated. The regional split is the only way to watch the ASF collapse arrive one region at a time.
PSA OpenSTAT — Philippine Statistics Authority statistical database
2,582,918 heads · 2026-Q2
+3.7%
Hogs slaughtered in registered slaughterhouses each quarter since 1987. Registered is the word doing the work here: backyard slaughter falls outside it, so this follows the commercial chain rather than what the country eats.
PSA OpenSTAT — Philippine Statistics Authority statistical database
975 1000 MT CWE · 2025
-2.5%
USDA's estimate of how much pork the Philippines produced, by market year since 1960. It disagrees with PSA's own production series and is published beside it rather than reconciled with it: the two use different methods and this archive does not average disagreeing sources.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
708 1000 MT CWE · 2025
+23.1%
How much pork USDA reckons the Philippines brought in, by market year since 1960. This is the number the import debate is actually about, and it is far larger relative to supply than PSA's import dependency ratio suggests — different methods, both named here, neither averaged.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
2 1000 MT CWE · 2025
0.0%
Philippine pork leaving the country, by market year since 1960. Nil or nearly nil in most of them, and USDA prints a literal zero both for a quantity that rounds to nothing and for one it does not estimate at all — every such row here says which it is not.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
1,687 1000 MT CWE · 2025
+8.5%
How much pork USDA reckons the country ate, by market year since 1960. Production plus imports minus exports and the stock change, which is why it moves with the import series rather than with the herd.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
92 1000 MT CWE · 2025
+24.3%
Cold-store pork on hand when the market year opened, by USDA market year. Small next to production in every year, and the reason the supply and use sides of the balance close.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
86 1000 MT CWE · 2025
-6.5%
Cold-store pork on hand when the market year closed. It is the following year's beginning stocks in every year USDA publishes both, which is one of the things worth checking about a balance sheet.
USDA FAS — Production, Supply and Distribution (PSD) livestock database
16 series · 1,636 rows
The wholesale ingredient prices a ration is costed from, read weekly out of the BAI report. Three of these columns have gone quiet at three different dates, and the dated blanks are in the files.
36.69 PHP/kg, premium grower/finisher tier · 2026-07-12
What a sack of premium grower/finisher feed costs per kilo, from dated retailer listings. One retailer on one day, and labelled as such: nobody publishes a national finished-feed price, and a trend line through a single point is not a trend.
Agrilife PH
24.75 PHP/kg, yellow corn, wholesale · as of 2026-07-13
0.0%
The single biggest line in a hog ration and the one that moves a sack price fastest. Weekly since January 2023, and the only official Philippine corn price published more often than monthly.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
27.10 PHP/kg, soybean meal (Argentina), wholesale · as of 2025-10-20
0.0%
Argentine soybean meal, the cheaper of the two imported protein sources BAI tracked. BAI stopped printing this column after 2025-10-20 and has printed N/A every week since. The readings and the dated nulls that follow them are both kept, so the week it went quiet is itself citable.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
46.40 PHP/kg, soybean meal (full fat), wholesale · as of 2026-07-13
0.0%
Full-fat soya, which carries its own oil and so prices well above the extracted meals beside it. Weekly and unbroken since January 2023.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
29.38 PHP/kg, soybean meal (US), wholesale · as of 2026-07-13
+21.2%
US soybean meal, the protein benchmark most Philippine formulations are actually costed against. The one to watch beside yellow corn: between them they are most of the cost of a grower ration.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
21.25 PHP/kg, feedwheat, wholesale · as of 2026-07-13
0.0%
Feed-grade wheat, the usual partial substitute when corn runs dear. Weekly and unbroken since January 2023, which makes the corn-to-wheat spread computable for every week of the series.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
18.50 PHP/kg, feedwheat (other origins), wholesale · as of 2024-08-12
0.0%
A second feedwheat line BAI printed for other origins. It stopped after 2024-08-12 — the first of the three ingredient columns to go quiet, and the longest dead. Its readings stand, followed by dated nulls carrying the report id that printed each blank.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
93.00 PHP/kg, local fishmeal, wholesale · as of 2026-05-18
+24.0%
Domestically produced fishmeal, the dearest protein in the table. It is also the only series in this archive whose source contradicts itself: on thirteen weeks BAI printed two different This Week prices for it inside one document, and every one of those weeks is a dated null naming each figure and the column that carried it. It stopped after 2026-05-18.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
106.00 PHP/kg, coconut oil, wholesale · as of 2026-07-13
+8.7%
Coconut oil as an energy source, and the most volatile line in the table by a wide margin — it has doubled and halved inside the life of this series. A ration that leans on it is exposed to the copra market rather than the grain market.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
14.60 PHP/kg, copra meal, wholesale · as of 2026-07-13
-0.7%
The cheapest protein source BAI tracks and the one most Philippine backyard rations actually use, because it is local. Weekly and unbroken since January 2023.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
82.80 PHP/kg, crude palm oil, wholesale · as of 2026-07-13
-0.4%
Crude palm oil as a ration energy source. READ THE NULLS BEFORE THE FIGURES: BAI printed no price in this column between 2024-11-04 and 2026-01-05, then resumed. A staleness warning on this series has fired before and is not evidence of a parser fault. Two further weeks were printed as a zero peso price, which is a blank rendered as a number and is recorded as no figure.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
25.75 PHP/kg, DDGS, wholesale · as of 2026-07-13
+1.0%
Distillers dried grains with solubles, the ethanol by-product used as a mid-protein energy filler. Weekly and unbroken since January 2023.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
17.85 PHP/kg, molasses, wholesale · as of 2026-07-13
+0.3%
Cane molasses, used for palatability and as a binder rather than as a major energy contributor. Weekly and unbroken since January 2023, and the steadiest line in the table.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
87.30 PHP/kg, palm olein, wholesale · as of 2026-07-13
0.0%
The refined fraction of palm oil, priced above the crude line beside it. Weekly and unbroken since January 2023 — which is what makes the crude-versus-refined spread readable even across the fourteen months the crude column was blank.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
22.30 PHP/kg, pollard (hard), wholesale · as of 2026-07-13
0.0%
Hard wheat pollard, the mill by-product that competes directly with rice bran for the same place in a backyard ration. Weekly and unbroken since January 2023.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
21.50 PHP/kg, rice bran D1, wholesale · as of 2026-07-13
0.0%
First-class rice bran, the grade most smallholders buy and the one whose price a raiser is most likely to check personally. The widest range in the table relative to its own level: it has more than doubled from trough to peak.
BAI — Livestock, Poultry and Feed Ingredients weekly price monitoring
3 series · 2,577 rows
Maize and soybean meal on the world market, and the peso-dollar rate the dollar-denominated series convert at. What a Philippine formulator is exposed to before anything lands.
213.40 USD/mt, world reference price · 2026-07
+8.3%
The world price of corn, monthly since 1960, from the World Bank's Pink Sheet. It is the number a Philippine feed miller's import cost is negotiated against, and it moves months before a sack price here does. Nominal US dollars — no inflation adjustment, and no peso conversion.
World Bank Commodity Price Data (“the Pink Sheet”) — monthly historical workbook
398.00 USD/mt, world reference price · 2026-07
-5.0%
The world price of the protein half of a hog ration, monthly since 1960, from the World Bank's Pink Sheet. BAI's Philippine soybean meal columns are landed prices quoted against this one. Nominal US dollars, not converted to pesos.
World Bank Commodity Price Data (“the Pink Sheet”) — monthly historical workbook
61.60 PHP/USD, monthly average · 2026-07
+0.6%
What a US dollar cost in pesos, month by month since 1945, from BSP's own reference-rate table. It is the rate the archive's dollar-denominated feed prices convert at, and it is BSP's figure rather than the IMF's — the two disagree badly through the decontrol years and this archive does not splice sources together.
Bangko Sentral ng Pilipinas — peso-dollar reference rate
1 series · 3 rows
The national active-barangay count, which is the reason pork prices move in this country and which nobody has published since May 2026.
8 active barangays · 2026-05-08
-70.4%
The national count of barangays with active African swine fever, as BAI reported it. The source stopped — no national figure has been published since 8 May 2026 — so the count is left where it stopped rather than carried forward. Individual outbreaks are still reported and are recorded separately.
Bureau of Animal Industry
An event has no value and no unit, so every field that makes an observation checkable is meaningless on it. These three are kept in a separate catalogue for that reason: no cards, no pages, no downloads, and deliberately absent from the machine-readable catalogue above.
4 records · outbreaks · hand-entered
Individual outbreaks as newsrooms and agencies reported them, hand-entered with the report behind each one. It exists because the national count died, and it is not a replacement for it: a list of outbreaks somebody wrote about is a floor, and a count is a count.
4,318 records · outbreaks · machine-retrieved
What the Philippines itself notified to WOAH, whose register is the official record. It stops before the hand-entered outbreaks begin, so nothing here supersedes anything there — different publisher, different method, no overlap.
7 records · orders · hand-entered
Orders, not outcomes. A price ceiling is what was decreed; whether anyone charged it is a farmgate series, and the two must never be read as one.
05 · What is missing
A series that goes quiet does not quietly go quiet. Each of these is a written, dated record of a number nobody publishes, and every series it silences links straight to it.
G-3
How much African swine fever there is right now
2 series · last checked 2026-08-15
G-4
BAI stopped printing its own farmgate prices, one column at a time
2 series · last checked 2026-08-15
G-6
BAI stopped printing its Argentine soybean meal price
1 series · last checked 2026-08-20
G-7
BAI stopped printing its feedwheat price for other origins
1 series · last checked 2026-08-20
G-8
BAI stopped printing its local fishmeal price, after contradicting itself about it
1 series · last checked 2026-08-20
G-9
PSA retired the backyard and commercial swine inventory split after 2020
3 series · last checked 2026-08-21
Counted from: the `gap_id` and `publishedAbsence` fields of the registry, resolved against the published records. A gap the registry stops citing leaves this section on its own.
06 · Take the whole thing
One long-format Parquet, every series stacked under a single schema, 237 KB. Withdrawn rows are excluded, so an average over it is correct without filtering. The unit is on every row rather than in the filename, because these series are in different units and grouping by series without reading the unit gives an answer the file itself shows you is wrong.
There used to be a stacked CSV here as well. It was 20.7 MB of exactly these rows, and
text/csv is not a type this CDN compresses, so it was
20.7 MB on the wire every time somebody clicked it. It is gone rather than hidden: nothing
it held has been withheld, and every row is still in the 4 files each series
publishes below.
| File | Format | What is in it |
|---|---|---|
| <series> .csv | text/csv | Published rows only, one row per period. Retracted rows are excluded, so an average taken over this file is correct without filtering. Null values are the token NA. |
| full/<series> .csv | text/csv | Every row this archive has ever held, including withdrawn ones, with a status column and the retraction date and reason. Use this to audit the archive, not to compute on. |
| <series> .json | application/json | The same rows as the full CSV, as JSON, with the coverage summary and the licence. |
| <series> .parquet | application/vnd.apache.parquet | The published rows with types: the numeric columns are doubles and a missing figure is a real null rather than the NA token a CSV has to write. |
Same strings as: the `distribution` blocks in every series page's Dataset markup and in `datapackage.json`. They are one array in `_catalog.ts`, so a page and a descriptor cannot describe the same file differently.
07 · For a machine
All three are generated from the same registry as the files, on every build, and every URL and byte size in them is asserted against the built output before a deploy.
The page on the file formats has the URL shapes, the column dictionary and worked snippets, if you are going to script against this.
08 · Using it
Free under CC BY 4.0 — use it, republish it, build on it, just say where it came from. The figures themselves are official statistics from PSA and BAI, which are public; what this archive adds is the retrieval, the provenance on every row, and the refusals. There is a page on the file formats if you are going to script against it, and a page on what is missing — the numbers nobody publishes, and what we checked before saying so.
Cite the archive as a whole
Baboy PH state-of-the-industry archive. CC BY 4.0. https://baboyph.com/data
The attribution is already inside the files. The citation line is the last column of every row rather than a header at the top, because the unit of loss is the row — somebody pastes four lines into a group chat, and a file-level notice does not survive that. Leave the column where it is and you have complied.
A single series has its own citation line on its own page, pointing at that page. There is no `/data/baboyph-industry` — the stacked Parquet is a repackaging, not a sixtieth series. Everything outside /data — the articles, the breed guides, the calculators — is under different terms: quote it with a credit and a link, just do not republish it whole.