Hi all,

Thanks very much for the responses.

I’ve created a wiki page that lists all of the ways in which I am identifying semi-automated tool edits. One thing that I’ve included there is a list of the regular expressions that I believe match tool edits from revision comments. Here is the page: https://meta.wikimedia.org/wiki/Research:Understanding_Wikidata%27s_Value/semi-automated_tool_edit_indicators.

I’ve also noticed the following interesting phrase within Wikidata revision comments: “Updating citation graph”. It appears that contiguous revisions with this phrase occur, indicating that this could be produced by a bot or some other form of automated editing occurring quickly. Anyone know what is producing these revisions? Is it a bot or tool possibly?

Thanks,
Andrew


On Oct 5, 2017, at 2:28 AM, Magnus Manske <magnusmanske@googlemail.com> wrote:

Also, [[User:QuickStatementsBot]] acts as an agent for edits submitted through the QuickStatements tool, as well as some of my automated bot edits ("...invoked by...").

On Thu, Oct 5, 2017 at 12:55 AM Tilman Bayer <tbayer@wikimedia.org> wrote:
Hi Andrew,

I would also look at edit tags [1], which capture many edits from
OAuth-based tools that don't leave an identifying word in the edit
summary field (example: reCh [2]).
You can query these using the change tag table [3].

[1] https://www.wikidata.org/wiki/Special:Tags (may take some time to load)

[2] e.g.
https://www.wikidata.org/w/index.php?limit=50&title=Special%3AContributions&contribs=user&target=%D4%B1%D5%B7%D5%A2%D5%B8%D5%BF&namespace=&tagfilter=&year=2017&month=2

[3] https://www.mediawiki.org/wiki/Manual:Change_tag_table

On Thu, Sep 28, 2017 at 7:34 PM, Andrew Hall <hall1467@umn.edu> wrote:
> Hello,
>
> I’m doing some analyses in which I want to identify Wikidata edits done via
> editing tools (e.g. via QuickStatements, etc…). To identify these edits,
> I've first flagged and removed bot edits and then I’ve generated a list of
> the 1000 most popular revision comment words (ignoring case and some
> punctuation characters as part of this process). Within this list of words,
> I've identified 15 words that I believe indicate tool edits. I’ve included
> these 15 words below.
>
> Does anyone know of tool edits that would be missed if I search for
> revisions that contain one of these 15 words in their comments? Put another
> way, are there editing tools not listed below? If so, can I identify edits
> from those tools from revision comments?
>
> #quickstatements
> #petscan
> #autolist2
> autoedit
> nameguzzler
> labellister
> #itemcreator
> #dragrefjs
> [[useryms/lc|lcjs]]
> #wikidatagame
> [[wikidataprimary
> #mix'n'match
> mix'n'match
> #distributedgame
> [[userjitrixis/nameguzzlerjs|nameguzzler]]
>
>
> Thanks in advance,
> Andrew Hall
>
> _______________________________________________
> Wikidata mailing list
> Wikidata@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>



--
Tilman Bayer
Senior Analyst
Wikimedia Foundation
IRC (Freenode): HaeB

_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata
_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata