Relevant materials:
relevance: Unique Devices by Region
Content by geo: Evolution of regional content
Content by gender: Knowledge Gaps Plots - Queering Wikipedia

@CMyrick-WMF clarified her request related to Relevance metrics. The original request has been updated.

Decision record for the "gender diverse" classification :

Summary
For Core Annual Plan Metrics Content metric, a classification schema was needed for the gender categories required to measure % new quality articles about women and gender-diverse people. A classification schema has been finalized and is being documented here.

Status
Classification finalized. (See "Decision" below).

Decision-making process

• Re. the name of the category as "Gender Diverse":
  • Discussions among metrics working group: decision made to name category "Gender Diverse"
• Re. the schema for classifying "Gender Diverse"
  • Discussions between Miriam and Queering Wikipedia community re. classification schema.
  • Discussions among Research & Decision Science staff (Kate, Miriam, Omari, Maya, Hamid, Caroline) and Community Growth staff (Becky), regarding the classification schema.
  • Determined alignment between Queering Wikipedia feedback, Community Growth methodology, Research methodology (especially Community Insights methodology).

Context and problem statement

For the 2023-2024 Core Annual Plan Metrics, the Content metric includes "% new quality articles about women and gender-diverse people." Because gender identities can be classified in different ways, and because various gender categorizations have been used in Foundation work, a schema was needed to determine how to classify "men", "women, and "gender-diverse" people for Core Annual Plan Metrics calculations.

Data for calculating the "% new quality articles about women and gender-diverse people" come from the Knowlege Gaps (Gender Gap) dataset. In this dataset, gender class is determined by the Wikidata property P21 (sex or gender). The class can take ~40 different values, as listed in this file. As such, a schema is needed for bucketing those classes into "men", "women", and "gender-diverse" for the Core Annual Plan Metrics content metric.

Risks/Caveats
We may need to revisit this schema as new gender classifications are entered into Wikidata.

Decision
Classification schema for "gender diverse":

--- Note: Under this schema, "men" refers to non-gender-diverse-identifying males and/or men, and "women" refers to non-gender-diverse-identifying females and/or women. As such, "men" excludes gender-diverse-identifying men, e.g. transgender men, and "women" excludes gender-diverse-identifying women, e.g. transgender women.

CASE
    WHEN (category IN ('male', 'cisgender male')) THEN 'Men'
    WHEN (category IN ('female', 'cisgender female')) THEN 'Women'
    ELSE 'Gender_Diverse'
END as gender

Classification schema for "gender diverse" (R version):

# Note: Under this schema, "men" refers to non-gender-diverse-identifying males and/or men, and "women" refers to non-gender-diverse-identifying females and/or women. As such, "men" excludes gender-diverse-identifying men, e.g. transgender men, and "women" excludes gender-diverse-identifying women, e.g. transgender women.

gendata$gender3category[gendata$category!="male" & 
                        gendata$category!="cisgender male" & 
                        gendata$category!="female" & 
                        gendata$category!="cisgender female"] <- "Gender_Diverse"

gendata$gender3category[gendata$category=="male" | 
                        gendata$category=="cisgender male"] <- "Men"

gendata$gender3category[gendata$category=="female" | 
                        gendata$category=="cisgender female"] <- "Women"

gendata$gender3category <- factor(gendata$gender3category, 
                                   levels = c("Men", "Women", "Gender_Diverse") )

Q1 22-23 baseline data was provided. (Thank you!)

Assigning back to @Hghani for Q1 23-24 data. (Note: July and Aug only, since Sep not yet available)

@CMyrick-WMF do you mind opening a new task for the Q1 data? since the original request was completed within the due date.

In T346160#9221434, @Mayakp.wiki wrote:

@CMyrick-WMF do you mind opening a new task for the Q1 data? since the original request was completed within the due date.

I'll be sure to create a new ticket for Q2 data, but I'll keep this one open until the Q1 request here is complete.

Updated ticket description:

• Added link to corresponding ticket (T346262)
• Added "Q1" in data being requested, for clarify, and to differentiate from future tickets where I'll be requesting breakdowns for subsequent quarters.

Knowledge gaps snapshots to use for Content metic: https://analytics.wikimedia.org/published/datasets/knowledge_gaps/content_gaps/

• Use the "2023-07" snapshot for July data
• User the "2023-08" snapshot for August data

Movement insights decided on the following: 1) back fill July and August 2023 underrepresented regions data with the October 6 snapshot as that has not been reported. Going forward each new month will appended to this snapshot to maintain historical consistency. 2) Since the gender classification needs to be recalculated using the schema provided by @CMyrick-WMF the October 6 snapshot was used with the up-to-date schema to generate all historical and current data. Going forward each new month will appended to these snapshots to maintain historical consistency.

Data provided. Thank you!

domain_wiki_devices_2022_2023.tsv

• Contains unique devices (Relevance Metric) calculations for Q1 2022-23 and Q1 2023-24, at the level of domain (e.g., en.wikimedia.org, en.m.wikimedia.org) and wmf region
• Q1 2022-23 data includes July, August, and September
• Q1 2023-24 data includes July, August, and September
• Cannot add domain numbers together because duplicates exist between mobile and desktop domains

project_family_region_unique_2022_2023.tsv

• Contains unique devices (Relevance Metric) calculations for Q1 2022-23 and Q1 2023-24, at the level of wmf region; aggregated all wikipedias per region
• Q1 2022-24 data includes July, August, and September
• Q1 2023-24 data includes July, August, and September

geo_quality_articles_wiki_2022.csv

• Contains article quality %s (Content Metric) calculations for new quality articles for Q1 2022-23, per geographic category
• Geographic categories represent the region that the content of the article is associated with, per wikidata (e.g., Napoleon article = ”Northern & Western Europe” geography)
• Q1 2022-23 data shows average of July, August, and September
  • Average = SUM(new quality articles about the region in July, August, September) / SUM(new quality articles about all regions, including unclassed, in July, August, September) * 100

geo_quality_articles_wiki_2023.csv

• Contains article quality %s (Content Metric) calculations for new quality articles for Q1 2023-24, per geographic category
• Geographic categories represent the region that the content of the article is associated with, per wikidata (e.g., Napoleon article = ”Northern & Western Europe” geography)
• Q1 2023-24 data shows average of July and August only
  • Average = SUM(new quality articles about the region in July and August) / SUM(new quality articles about all regions, including unclassed, in July and August) * 100

quality_gender_updated_2022.csv

• Contains article quality %s (Content Metric) calculations for new quality biographies for Q1 2022-23, per gender category
• Gender categories represent the gender of the person the biography is about (e.g., Napoleon article = “men” gender category)
• Q1 2022-23 data shows average of July, August, and September
  • Average = SUM(new quality biographies about women and gender-diverse people in July, August, September) / SUM(new quality biographies about all genders, in July, August, September) * 100

quality_gender_updated_2023.csv

• Contains article quality %s (Content Metric) calculations for new quality biographies for Q1 2023-24, per gender category
• Gender categories represent the gender of the person the biography is about (e.g., Napoleon article = “men” gender category)
• Q1 2023-24 data shows average of July and August only
  • Average = SUM(new quality biographies about women and gender-diverse people in July and August) / SUM(new quality biographies about all genders, in July and August) * 100