Event Timeline
Hey, would it be possible to run reports like this on a regular basis? In the Wikipedia Library, we are manually tracking link metrics, and don't have a good way to compare growth of partnerships, with items like mean/average growth of partners, and historical trends.
Moreover, it would be awesome to have some type of historical reports available that visual that data in the realm of something like https://tools.wmflabs.org/glamtools/baglama2/
@Sadads It would be possible, but it's not convenient at the moment. I wrote the tools to do this for a one-off project so it's a bit clunky, and not ready for regular reports.
I'm focusing on other projects now, but if anyone is interested, https://github.com/gpaumier/refsdomains#refsdomains explains how to run the script; anyone should feel free to fork it, or submit pull requests to make it more robust and automated.
@gpaumier thanks for letting us know, sorry for taking so long! We don't have anyone to run this right now, but are going to explore if one of our volunteers is interested in helping. It would be great if we could send someone to consult with you, at some point. It would also be great to run a historical test of this data: such as figuring out 1 year back or 3 years back, what the rankings were in citations.
@gpaumier I talked with one of our volunteer, says script requires a bit to large of download and wrangling on his end: https://en.wikipedia.org/wiki/User_talk:Johnuniq#Another_Query_for_TWL . Is there a way it could be done without huge download?
So I did a quick visual analysis for T115119 looking for suffixes with more than one parts, I also included other top level suffixes; here is my totally unscientific list, probably missed a few:
.com.au
.au
.govt.nz
.info
.gov
.com
.org
.edu
.net
.co.uk
.fi
.us
.ie
.co.nz
.ac.uk
.co.in
.ca
.de
.si
.ch
.gov.in
.gov.au
.mil
.tv
.jp
.gc.ca
.com.cn
.co.jp
.co.kr
.it
.fr
.com.ar
.edu.au
.br
.es
.com.pk
.com.mx
.com.au
.gob.mx
gov.bf
.eu
.com.hk
.ph
.no
.pt
.int
.se
.lk
.ru
.com.cn
.it
.go.th
.tr
.com.ua
.com.br
.gov.lk
.co.in
.gov.au
.org.uk
.org.eg
.no
.co.il
.co.kr
.hr
.int
.nl
.nam
.pl
.co.uk
gc.ca
.co.za
.govt.nz
.net.au
.gr
.az
.gov.uk
.co.jp
.cl
I have submitted a grant proposal to perform this type of analysis on a more systematic basis: https://meta.wikimedia.org/wiki/Grants:Project/Harej/Librarybase:_an_online_reference_library
While we wait for James' grant proposal to succeed, any chance of updating this as a one-off please?
And looking at the most-used sites with no tailored Zotero translator?
While we wait for James' grant proposal to succeed, any chance of updating this as a one-off please?
And looking at the most-used sites with no tailored Zotero translator?
Maybe we could also combine domains; e.g. http://www.biodiversitylibrary.org and http://biodiversitylibrary.org are really equivalent to each other; as are http://books.google.com and http://books.google.co.uk
As a followup - we used this list to generate some stats and research for Wikicite 2017. Slides and raw data below:
Citoid performance for news article citations
https://docs.google.com/presentation/d/1HI1FMZQ1Wkg15d1ny8Qynfq4JkvPA_C46r7cOVm2xDM/edit
Public Citoid performance for news 2017
https://docs.google.com/spreadsheets/d/1kNHSVKq5qZr6_3-05UFtXF2RA5sn3HdUUgL0nC57XPk/edit