@CBogen, @Ramsey-WMF and @Abit : head's up that this task isn't likely to go anywhere during the current quarter given that MediaSearch and Computer-Aided-Tagging are priorities. Keeping it in the "Upcoming Quarter" column on the PA board for now, to be revisited in Q2.

The goal of this task is unclear, which makes it difficult to prioritize it. If we are to undertake this work, we need to know more about how building this dashboard will help inform the org or drive decisions. This is partly because there's already tracking of the number of files with at least one structured data element (T238878#5692516, productionized in T239565, see also T247101), which arguably tracks the growth of SDC. For now, I'll move this to the upcoming quarter for a future revisit.

In T252443#6770723, @nettrom_WMF wrote:

The goal of this task is unclear, which makes it difficult to prioritize it. If we are to undertake this work, we need to know more about how building this dashboard will help inform the org or drive decisions. This is partly because there's already tracking of the number of files with at least one structured data element (T238878#5692516, productionized in T239565, see also T247101), which arguably tracks the growth of SDC. For now, I'll move this to the upcoming quarter for a future revisit.

The goal of this ticket is to have a regularly updated visual dashboard (in superset or something like it?) where a PM or other interested party can easily see the number of files with at least one structured data element, as well as the other data in the notebooks listed in the description, e.g.: Media files containing structured fields in non-English languages; Number of files with non-English captions; Number of files with English captions; Number of files that had captions added; How quickly after creation does non-English captions get added?; and Time to edit for SDC; Comparison of number of SDC and non-SDC edits.

I'll update the description to add this.

@CBogen and I discussed this in our sync meeting today, as reflected in the updated task description. I've added T258834 as a subtask of this one, because having the SDC data queryable in the Data Lake will help us answer key questions about the richness of it.

The number of media files with M-entities has been measured monthly for the past year and half and you can get that from commonswiki_mediainfo_slots.tsv on analytics.wikimedia.org (phab task for that was T239565). Looks like there’s also a Grafana dashboard that tracks the number of pages on Commons with Wikidata entity usage (found through T238878). The number is considerably higher than the number of media files, as one would expect.

The draft dashboard have the following monthly metrics available from Jun 2021 - Dec 2021:

Captions

• Number of files with captions added monthly
• Number of files with non-English/English captions added monthly
• Number of files with both captions added monthly
• How quickly after creation do non-English captions get added?

SDC edits

• Number of SDC and non-SDC edits monthly
• Time to edit for SDC

For caption-related metrics, we use mediawiki_history table to calculate the captions added per month. For a total number of files with captions, we decide to calculate it after the entity table is created.

For SDC edits metrics, I chose to ignore whether these edits were reverted as I've previously found that reverts of these are rare (as are removals). For wikitext edits, they have to also not be a revert nor be reverted (within the common timeframe of 48 hours) as those are either reinstating an old version of the page or an unproductive edit.
In addition, these are all non-bot edits, and we only count edits to files that have not been deleted.

Following metrics were added to the draft dashboard.

• Number (%) of files with at least one structured data element
• Number of files with license
• Number of files with depicts
• Number of files with captions (en vs. non-en vs. both)

The metrics are from commons_entity table. We have the data back to Dec 2021.

Question: the "SDC edits" tab mentions "only look[ing] at non-bot edits to files".
MediaInfo submits all edits with a bot: 1 flag via the API (code for statements, code for captions) (the reason is simply "because that's how Wikidata does it")
Does that mean that those edits (via the regular SDC UI) are not included in this data, or does that param have no impact on how a "bot edit" is determined?

In T252443#7873533, @matthiasmullie wrote:

Question: the "SDC edits" tab mentions "only look[ing] at non-bot edits to files".
MediaInfo submits all edits with a bot: 1 flag via the API (code for statements, code for captions) (the reason is simply "because that's how Wikidata does it")
Does that mean that those edits (via the regular SDC UI) are not included in this data, or does that param have no impact on how a "bot edit" is determined?

For SDC edits, we used edit comments in mediawiki_history to identify structured data edits (notebook). And the way we flagged bot is based on event_user_is_bot_by in mediawiki_history.