WMDE Analytics Request

This task was generated using the WMDE Analytics request form. Please use the task templates linked on our project page to create tasks for the team. Thank you!

Why (Context & Decision)

We need to better understand how the different segments of Wikidata's Items develop over time. This is important for us to understand as part of our work to combat promotional Item creation as well as for discussions about the future of scholarly articles.
We'd like to understand if specific segments of our content are growing faster than others as well as what the current percentage of different segments are.

What (Scope & Output)

We will look at "instance of" values for every Item and assign it to a bucket. The buckets to start are:

• human
• scholarly article
• astronomical object (instances of star, galaxy, infared source)
• Wikimedia-internal entity
• taxon
• gene + protein
• chemical compound
• company
• name
• art and media
• sport (for example this would not include human athletes, as they would go within "humans" category)
• human settlement
• other

Some of these like Wikimedia-internal entity will combine several instance-of-values. We will probably not be able to follow the class hierarchy so let's accept that we will have a hand-curated list for the main ones.

We will look into entities that fall into "other" together and see if there are other larger categories hidden in there or if some of the Items that fall into that category should move to an existing category.

• Output: the distribution of sub classes / categories within OTHER to see what can be further grouped before final DAG and output is built

By When (Timing & Priority)

Provide a clear deadline, any key milestones (e.g. launch, leadership review), and note if timing is flexible or fixed.

End of September / Mid-October

Other useful info

• Adam did a lot of work in this blog and described the methodology we can follow

---

Information below this point is filled out by the analyst.

Sub Tasks

A breakdown of the steps to complete this task.

One Time Data Investigation

• Define application / usage of the investigation output: is it clear the question we are trying to answer and how this will help inform the stakeholder(s)
  • Thorough investigation of PIDs for the groups
  • An investigation of "Other", and how this can be broken down / the definition of the categories
  • Use the classes / class hierarchy provided in the ticket, then look at the "other" category to see whether these can be broken down further
• Check in between the Analyst and Product on the groupings of the entities based on PID
• Check in between the Analyst and Product on the output of the Other investigation and how the PID groupings need to change based on what was found in Other
• Define investigation output with stakeholder(s)
• Define the timespan and sample size of investigation
• Write queries and/or code for investigation
• Deliver results of investigation

Finishing Up

• Required documentation is written
• The task is moved to Delivered with the stakeholder being assigned and CCed in a comment

Data

The tables that will be referenced in this task and the sample sizes from them that will be used.

• link_to_table