@Manuel

Initial check for Tag all edits made via Wikibase View and Wikibase Client T236893:

• data-bridge: used 279 times;
• client-linkitem-change: used 58740 times;
• client-automatic-update: used 58367 times;
• wikidata-ui: used 902690 times;
• termbox: used 6384 times.

Checks were performed from the wikidatawiki.change_tag_def table.

Thx Goran! Could you please figure out where the problem is? The tags are already in use.

@Manuel

Thx Goran! Could you please figure out where the problem is? The tags are already in use.

The tags are in use indeed, but you must have seen T285459#7321218 before I had it edited.
My bad: on the first sight I did not spot the new tags in the table (attached). Please take a look again.

Thank you! Let's discuss how we could use these tags in combination with other inf (e.g isBot and the total number of edits).

@Manuel

Let's discuss how we could use these tags in combination with other inf (e.g isBot and the total number of edits).

I am currently working on this from SQL, because the august 2021 snapshot of wmf.mediawiki_history is not yet produced in our Data Lake.

It might take some time to get to the results since a rather tricky join (change_tag_def --> change_tag --> revision_actor_temp --> actor --> user) needs to be performed in order to find out how many of the new tags were used by who in wikidatawiki since 20210801000000.

I will not be using the user.user_editcount since it is a rough approximation of the number of revisions made on behalf of a user; rather join user on revision_actor_temp --> actor.
Also, I will need to filter out bots manually in this approach, and I have already created two fields in that respect (botByName, botByGroup), similarly to what is done in wmf.mediawiki_history in the Data Lake.

To put it in a nutshell:

SELECT actor.actor_user AS user_id, 
                change_tag_def.ctd_name as rev_tag, 
                COUNT(change_tag_def.ctd_name) as rev_tag_count 
        FROM actor 
        LEFT JOIN revision_actor_temp ON (actor.actor_id = revision_actor_temp.revactor_actor)
        LEFT JOIN change_tag ON (revision_actor_temp.revactor_rev = change_tag.ct_rev_id) 
        LEFT JOIN change_tag_def ON (change_tag.ct_tag_id = change_tag_def.ctd_id)
        WHERE revision_actor_temp.revactor_timestamp > 20210801000000 
        GROUP BY actor.actor_user, change_tag_def.ctd_name;

FWIW a week-based quarry on semi-automated edits from the most popular tools (manually selected based on another query) is here: https://quarry.wmcloud.org/query/58473
Results:
OAuth CID: 1776 869604 aka QS
wikidata-ui 403984
client-linkitem-change 30437
client-automatic-update 20785
WikibaseJS-cli 20100
OAuth CID: 1740 13024 aka Author-Disambiguator 2.0
openrefine-3.4 6980
apps-suggested-edits 6628
OAuth CID: 1768 6002 aka Mix'n match 1.0
termbox 3422
openrefine-3.3 2324
data-bridge 1

According to @Lydia_Pintscher the lexeme edits are not getting the wikidata-ui tag yet.

Statistics are:
manual edits on wikidata.org based on a week of data make up (403984+3422+6628+1)/(869604+20100+6980+6002+2324+13024)=414035/918034=0.45 ~45% of the total edits

@So9q Thank you very much for this insight!

Thank you, Dennis! And nice to meet you in Alessandro's session!

Happy to be of service, and nice to meet you too. I liked the presentation a lot. I feel like doing a ph.D myself on knowledge graphs like Wikidata 😆

So about the statistics, I'm actually surprised the percentage is so high. It will be interesting to follow this over time.