Korean pop culture, in numbers

The data behind these pages

Nothing on this site is written from a press release. Every figure comes out of tables we build and keep — 15,954 rows across 8 files. This page says what is in them, and it says what is missing from them before it says anything else.

What is in it

Tables8 files, one row per thing counted
Rows15,954 across all of them
FormatUTF-8 CSV, plus a column dictionary that says what every column means and what a blank cell means
Covers2021-07-04 to 2026-08-23 for the chart tables; the last 30 days for daily attention
Refreshedweekly, as Netflix publishes · with each new charting title · daily · monthly, as Wikidata fills · yearly, as companies file · the day we correct anything · whenever a gap opens or closes · with every column we add
Also includedThe corrections record and the coverage file, so you can see what we changed and what is missing before you build on it

Per table, the refresh is what the source publishes, not how often we feel like looking: Korean titles on Netflix in Southeast Asia — weekly, as Netflix publishes; Who appears in what — with each new charting title; K-pop attention — daily; Korean entertainers, with their schools — monthly, as Wikidata fills; Listed content companies — yearly, as companies file; Everything we got wrong — the day we correct anything; What is missing — whenever a gap opens or closes; What every column means — with every column we add.

Two rulers, not one

The first question anyone should ask about a panel is what put a row in it. Ours has two answers.

The name. Netflix publishes no country of production, so a title enters because its English name matches a Korean work in Wikidata. That can fail in exactly one way — another country made something with the same name — so we measured how often it can. 149 of 421 titles carry a name a foreign work also carries.

Korea's own chart. A Korean title normally also plays in Korea, so we counted every title's weeks on Netflix's South Korea list. 75 of those 149 have never appeared on it, and 20 charted in exactly one country in the world. Those columns are in the table, per row, so you can sort by them yourself.

Neither ruler decides alone and we do not hide the disagreement. A Korean work can miss Korea's chart because it was released before this data begins, or was never streamed there. Nothing has been removed on the second ruler's strength alone — it produces a queue to read, and anything we do remove appears on the corrections page with the evidence.

What is missing

We would rather you learned this from us than from a spreadsheet at six in the evening.

The same list is in coverage.csv, in a form you can filter, with a column saying whether each gap can ever be filled.

Take the tables

Free to use, including commercially. Please cite as K Culture Wire and link back, so a reader can see how the number was built. An empty cell means we do not hold the value — it never means zero.

Three tables are not here, and we will say why. The chart figures on this site are counted from Netflix’s own weekly country lists. Publishing counts built from them is one thing; handing over a file that sits close to the source is another, and we have not been able to read the terms that govern the second. The Top 10 pages are drawn by script, so fetching them returns an empty document — we could not read the conditions as text. That is “we have not seen it”, not “there is nothing there”, and we will not treat those as the same thing. So the aggregate figures stay on the pages where you can read and cite them, and the near-source tables stay off until we can read the terms or get an answer in writing. If you need a figure that is on a page but not in a file, ask us.

If a figure looks wrong, tell us. We publish corrections with the old number still visible — every one we have made is here.

Came here looking for Netflix's own files? The two files most of this rests on are published by Netflix as tab-separated downloads, and they carry two traps that are written nowhere on them. We measured both files and wrote down what is in them, and what is missing from them.

How to get it

Email cs@klifedesign.net and say what you are trying to answer. We will tell you whether these tables can answer it — including when they cannot, which is the more useful reply.

What happens then. We reply within one working day, in Korean or English, with a plain answer to whether these tables cover your question. If they do, you get the files and the column dictionary; if they do not, we say what would have to exist and we do not sell you the near miss.

Two of these panels are ready to hand over today. The K-pop attention table and the listed company table rest only on Wikimedia, Wikidata, KOSIS and Korea's corporate filings, and we have checked those terms. They are bundled separately for exactly that reason, with the corrections record alongside them, and nothing in that bundle is derived from a streaming chart.

$39 for the bundle as it stands, or $19 a month for the refreshes and whatever tables we add. One licence covers one organisation — the terms are short.

The opening month is free. Everything above is open at no charge from 15 August to 14 September. That is the full price standing with the month opened, not a discount — we would rather you saw what it costs and then decided.

Why that price. Half of what we sell is marked empty, and we say so on the page before you buy. A price that ignores that would be a price for a table we do not have. Nothing comparable is sold, so we cannot read the market. When we cannot read it we start low and raise it, rather than start high and discount.

One table is waiting on someone else. We are not offering the chart-derived tables as a public download yet: the redistribution terms for that source are still being confirmed, and we check a licence before we publish, not after. The other tables do not wait on it, and the Netflix-free bundle is priced separately once we know what we are allowed to include.

What we wrote from this data

← Back to the front page