@Pginer-WMF I have consolidated all the metrics we are currently tracking across dashboards / reports and are good to have for the first version. Additionally, from this search, I have listed the open tickets and if we can addressed those with the first version of the dashboard. Please review and share your thoughts.

Listing down all the metrics for future reference (priority can be discussed and changed)

Machine translation service usage report (source: cx_corpora, cx_translations, & mediawiki_history; frequency: ad-hoc)

CX Abuse Filter events (source: wmf_product.cx_abuse_filter_daily)

CX funnel analysis (v1 - entry points) (frequency: ad-hoc)
Note: This is dependent on the instrumentation of CX event, it is better to build the entire after all the key events are instrumented.

In T366044#9840860, @KCVelaga_WMF wrote:

@Pginer-WMF I have consolidated all the metrics we are currently tracking across dashboards / reports and are good to have for the first version. Additionally, from this search, I have listed the open tickets and if we can addressed those with the first version of the dashboard. Please review and share your thoughts.

Thanks for the detailed evaluation. The adjustments proposed for merging the multiple dashboards/reports available make perfect sense.
Regarding, T228152, I think the current approach in the Key metrics dashboard will cover the main needs. In particular the "Monthly translations by wiki" graph has been useful to identify which wikis caused a spike of activity (e.g., a campaign in Bangla Wikipedia)

Another report that would be relevant to consider integrating is the Topic Diversity of Published Translations. As we provide support for topic-based suggestions(T113257), having an understanding about the topics associated with the published translations can be useful in this context.

Definitely not a blocker for the initial version. Sharing it since other reports were mentioned for future versions in T366044#9897587

I wanted to share a couple of proposed aspects to measure that have been surfaced in several conversations in the past as the impact of translations have been analyzed:

• Distribution of translated content across translators. We want to understand how are translations distributed across the users making them, and how much content do they translate. This will help to identify different groups of our audience and how representative of the whole audience they are. For example: a group of users creating many short translations, another making a few translations that cover most of the original article, etc.
• Impact on readers and other editors. For content created as a translation, it would be useful to know how many views does it generate and how often other editors build on top of the translation to continue editing (i.e., increasing the coverage on the topic by continuing where the translation left).

@Pginer-WMF these are definitely good points. But given the scope of the building the first version of the public dashboard, let's plan these as expansions during the upcoming quarters.

Distribution of translated content across translators. We want to understand how are translations distributed across the users making them, and how much content do they translate. This will help to identify different groups of our audience and how representative of the whole audience they are. For example: a group of users creating many short translations, another making a few translations that cover most of the original article, etc.

I am wondering if this will have much value being in a dashboard and monitoring continuously, vs. a one time analysis.

In T366044#10890512, @KCVelaga_WMF wrote:

@Pginer-WMF these are definitely good points. But given the scope of the building the first version of the public dashboard, let's plan these as expansions during the upcoming quarters.

Distribution of translated content across translators. We want to understand how are translations distributed across the users making them, and how much content do they translate. This will help to identify different groups of our audience and how representative of the whole audience they are. For example: a group of users creating many short translations, another making a few translations that cover most of the original article, etc.

I am wondering if this will have much value being in a dashboard and monitoring continuously, vs. a one time analysis.

Good question. I'd say that it depends on whether the plans for the team are focused towards affecting the aspect we measure. For example, if the goal is about encouraging people to make higher quality articles, it may be useful to see how the distribution of translated content per user evolves. Obviously, this also depends on the differences of efforts required for the measurements on each approach.

Maybe we can consider something along these lines when measuring the impact of future hypothesis if any is related.