Add publication stats page

[mjpost]

This is a draft PR, forked from #9179. It's in draft mode until we can get complete grob-id-based stats.

[github-actions]

Build successful. Some useful links:

• Complete site preview: https://preview.aclanthology.org/stats
• Potential changes of interest: 2016.gwc-1, 2020.acl-main, 2020.acl-demos, 2020.acl-srw, 2020.acl-tutorials, 2020.findings-emnlp, 2021.acl-long, 2021.acl-short, 2021.acl-srw, 2021.acl-demo, 2021.acl-tutorials, 2021.eancs-1, 2021.emnlp-main, 2021.emnlp-demo, 2021.emnlp-tutorials, 2022.coling-1, 2022.lrec-1, 2022.seretod-1, 2022.sigdial-1, 2022.wmt-1, 2023.acl-long, 2023.acl-short, 2023.acl-demo, 2023.acl-srw, 2023.acl-industry, 2023.acl-tutorials, 2023.ccl-1, 2023.ccl-2, 2023.ccl-3, 2023.ccl-4, 2023.conll-1, 2023.conll-babylm, 2023.cxgsnlp-1, 2023.dmr-1, 2023.emnlp-main, 2023.emnlp-tutorial, 2023.emnlp-demo, 2023.emnlp-industry, 2023.findings-eacl, 2023.findings-acl, 2023.findings-emnlp, 2023.findings-ijcnlp, 2023.gwc-1, 2023.law-1, 2023.sigmorphon-1, 2023.tlt-1, 2024.acl-long, 2024.acl-short, 2024.acl-demos, 2024.acl-srw, (plus 259 more...)

This preview will be removed when the branch is merged.

[nschneid]

A thought: could be interesting to plot a few things, organized in a tabbed display:

• number of authors by year of earliest paper
• number of authors by year of most recent paper
• number of authors by year of peak number of papers
  • calculated as: N(author,yr) = number of papers the author published that year; N_peak(author) = max_yr N(author,yr); yr_peak(author) = median({yr: N(author,yr)=N_peak(author)}), breaking ties in favor of the later year

Grouping by decade pre-2020s as those counts will be much lower.

Unrelated idea: a page that lists notable authors (with their photos) for browsing. Maybe ACL Fellows + LTA winners. Sorted by the author's peak publication year?

Also: list of top 100ish authors by longevity, defined as the number of years in which the author had at least 1 paper.

(Please no ranking by raw paper count. We don't want to reward authors who maximize paper count at the expense of quality.)

[nschneid]

Some longevity stats:

>>> sum(longevityC[x] for x in longevityC if x>=5)
11879
>>> sum(longevityC[x] for x in longevityC if x>=10)
3160
>>> sum(longevityC[x] for x in longevityC if x>=20)
590
>>> sum(longevityC[x] for x in longevityC if x>=30)
79
>>> sum(longevityC[x] for x in longevityC if x>=40)
5

I.e. 79 authors have published in at least 30 distinct years. They are:

• 43: Bonnie Webber, Jun’ichi Tsujii, Kathleen McKeown
• 42: Aravind Joshi
• 40: Jaime G. Carbonell
• 39: Chu-Ren Huang, Eduard Hovy, Eva Hajicova, Martha Palmer, Ralph Grishman
• 38: Julia Hirschberg
• 37: Hsin-Hsi Chen, Ted Briscoe
• 36: Jan Hajic, Mark Johnson, Mark Steedman, Owen Rambow
• 35: Alex Waibel, Diane Litman, James Pustejovsky, Yuji Matsumoto
• 34: Christian Boitet, Graeme Hirst, Ido Dagan, Ralph Weischedel, Walter Daelemans
• 33: Barbara Di Eugenio, Claire Cardie, Claire Gardent, Ellen Riloff, Eric Nyberg, Francis Bond, James Allen, Jason S. Chang, Kenneth Church, Marilyn Walker, Philip Resnik, Roberto Basili, Ruslan Mitkov, Sadao Kurohashi, Stuart M. Shieber, Teruko Mitamura, Yorick Wilks
• 32: Dan Jurafsky, Eiichiro Sumita, Massimo Poesio, Michael Strube, Satoshi Sekine
• 31: Alon Lavie, Carolyn Rose, Chin-Yew Lin, Christopher D. Manning, David Weir, David Yarowsky, Ehud Reiter, Manabu Okumura, Masaaki Nagata, Michael White, Rebecca J. Passonneau, Takehito Utsuro, Takenobu Tokunaga
• 30: Anoop Sarkar, Bonnie Dorr, Dan Roth, Eneko Agirre, German Rigau, Giorgio Satta, Hwee Tou Ng, Jason Eisner, Kevin Knight, Lori Levin, Manny Rayner, Michael Collins, Paola Merlo, Philippe Blache, Pierrette Bouillon, Richard Sproat, Steven Bird, Udo Hahn

[mjpost]

Great idea and interesting list.

I like the idea of a hall of fellows, but I think that's beyond the scope of this PR.

[nschneid]

Nice!

[nschneid]

I'm seeing a scrollbar for the tabs:

[nschneid]

Bar plots: Is it possible to put a vertical line or something to show where the scale switches from decades to single years?