Open data

Everything on this site is available as a file. No key, no rate limit, no attribution required — though a link back is always welcome, and telling me what you built is even better.

The files

conferences.json — one object per edition, with the

full record

conferences.csv — the same, flattened, for

spreadsheets and pandas

Both are regenerated whenever the site is rebuilt, which is roughly fortnightly.

What a row contains

One row per edition. The columns worth explaining:

paper_deadline is the date that actually applied; paper_deadline_announced

is the date first published, and paper_extended_days the difference. Keeping all three is the point of this dataset — nowhere else records what a deadline was.

timezone is empty when the organisers did not state one. It is not filled in

with a guess.

submitted and accepted are the overall figures. Where a conference

publishes only per-track numbers, these are their sum.

lat and lon come from a hand-built gazetteer and locate the city, not the

venue.

An empty field means the information is not recorded, which is not the same as the information being absent in reality. See about.

Calendars

If you want to follow deadlines rather than analyse them, the calendar feeds are more convenient.

Licence

CC BY-SA 4.0 — the same terms as Wikipedia. Use it for anything, including commercially, on two conditions: say where it came from, and publish anything you build on it under the same licence.

The second condition is the one that matters. This is a decade of somebody reading calls for papers and typing dates in, and share-alike means a site that copies it has to leave its own version equally open. That is a far better protection than forbidding commercial use, which would also block a university spin-off or a conference aggregator that happens to run ads.

Attribution: a link to <https://www.conferences-computer.science> and/or Étienne André is enough.

Calls for papers reproduced under the archive belong to their authors and are kept as they were published; the licence above covers the compilation, not them.

Data entirely hand-typed, one call for papers at a time. There is no scraper behind any of this: I read the announcement and type the dates in, in the gaps between other work, and with the hope it is useful to others. Organisers then move deadlines without telling anyone, calls get edited in place, and now and then the typing is simply wrong. Treat this as a well-meaning notebook rather than an authority, and check the conference’s own page before you submit anything you care about.