What projects do volunteers from Sub-Saharan African edit most?¶

Date: 17 June 2022 Task

Scratch Notes¶

Do not know first time editors currently only 1-4 grou? Any way possible in aggregated data? Might need to look at raw data
Find a quick way to limit project familty to wikipedia
There is also Geoeditors Edits Monthly which show the number of edits not distinct editors.
How are anons handled in the data?

Purpose¶

For the 2022-2023 fiscal year, the Editing Team is considering centering the work we will do around people A) who are contributing to a Wikipedia from sub-Saharan Africa and B) who have made fewer than 100 cumulative edits to Wikipedia.

Before moving forward with defining the audience in the ways "A)" and "B)" describe, we would like to understand:

What language Wikipedias the people who are contributing from sub-Saharan Africa and who have made fewer than 100 cumulative edits are most active with?

Knowing the above will help us determine what projects we will consider partnering most closely with in the 2022-2023 fiscal year.

On average, how many people from sub-Saharan Africa during a given month are making their first edit to Wikipedia as an account holder?

Knowing the above will help us determine how long we can expect the quantitative analyses we have planned to reach statistic significance.

Data caveats¶

This data comes from from geoeditors_monthly which is accessible via Superset.
Mediawiki_history joined with X for geo location data is needed to access details on first time editors. Query and results for that are noted below.
Reviewed data logged from 1 January 2022 data to date (17 June 2022)
Limited analysis to countries within Subsaharan African region as defined in https://docs.google.com/spreadsheets/d/1xPzo1KtspT2EaWbO_HfLHRvuk9BfbXsfwjkrOV_5m2M/edit#gid=741931431
Geoeditors Data https://superset.wikimedia.org/superset/dashboard/9/

References¶

Regional pageview and editor metrics
Irene’s recent work on SSA Campaigns activity T287715
Irene's superset dashboard
Sub-saharan Africa country names and codes

Data Notes:¶

Data comes from wmf.editors_daily and wmf.mediawiki_history table.
Data is limited to May 2022 as data from wmf.editors_daily is only retained for the last two calendar months.
This reflects editors who made only 1 edit on the wiki.

On average, how many people from sub-Saharan Africa during a given month are making their first edit to Wikipedia as an account holder?¶

In [29]:

shhh <- function(expr) suppressPackageStartupMessages(suppressWarnings(suppressMessages(expr)))
shhh({
    library(tidyverse); library(glue); library(lubridate); library(scales)
})

May 2022¶

In [56]:

query <-

"
SELECT
mwh.event_user_id as user_id,
CASE
    WHEN min(event_user_revision_count) = 1 THEN 'first_time_editor'
    WHEN min(event_user_revision_count) < 100 THEN 'under 100'
    WHEN (min(event_user_revision_count) >=100 AND min(event_user_revision_count <= 500)) THEN '100-500'
    ELSE 'over 500'
    END AS edit_count_group,
mwh.wiki_db AS project
FROM 
wmf.mediawiki_history mwh
INNER JOIN canonical_data.wikis
ON wiki_db = database_code and
  database_group ==  'wikipedia'
INNER JOIN
wmf.editors_daily ed
ON  mwh.event_user_id = CAST(ed.user_fingerprint_or_id AS BIGINT)
AND mwh.wiki_db = ed.wiki_db
WHERE
-- from sug-saharan Africa region
ed.country_code IN ('AO','BJ', 'BW','BF','BI','CM','CV','TD','KM','CI', 'DJ', 'GQ', 'ER', 'ET', 'GA','GM','GH',
'GN','GW','KE','LS','LR','MG','MW','ML','MU', 'MZ', 'NA', 'NE', 'NG', 'CD', 'RW', 'ST', 'SN', 'SC', 'SL', 'SO', 'ZA', 
'SD', 'TZ','CF', 'CD', 'TG', 'UG', 'ZM','ZW')
AND ed.month = '2022-05'
AND mwh.snapshot = '2022-05'
AND mwh.event_timestamp >= '2022-05-01'
AND mwh.event_timestamp <= '2022-06-01'
AND ed.user_is_anonymous = FALSE 
AND mwh.event_user_is_anonymous = FALSE
GROUP BY 
mwh.event_user_id,
mwh.wiki_db
"

In [57]:

first_time_editors_may <-  wmfdata::query_hive(query)

Don't forget to authenticate with Kerberos using kinit

In [62]:

first_time_editors_total_may <- first_time_editors_may %>%
    #filter(edit_count_group == 'first_time_editor')  %>%
    filter(project == 'enwiki') %>%
    summarise(n_users = n_distinct(user_id))

first_time_editors_total_may

A data.frame: 1 × 1
n_users
<int>
10534

April 2022¶

In [49]:

query <-

"
SELECT
mwh.event_user_id as user_id,
CASE
    WHEN min(event_user_revision_count) = 1 THEN 'first_time_editor'
    WHEN min(event_user_revision_count) < 100 THEN 'under 100'
    WHEN (min(event_user_revision_count) >=100 AND min(event_user_revision_count <= 500)) THEN '100-500'
    ELSE 'over 500'
    END AS edit_count_group,
mwh.wiki_db AS project
FROM 
wmf.mediawiki_history mwh
INNER JOIN canonical_data.wikis
ON wiki_db = database_code and
  database_group ==  'wikipedia'
INNER JOIN
wmf.editors_daily ed
ON  mwh.event_user_id = CAST(ed.user_fingerprint_or_id AS BIGINT)
AND mwh.wiki_db = ed.wiki_db
WHERE
ed.country_code IN ('AO','BJ', 'BW','BF','BI','CM','CV','TD','KM','CI', 'DJ', 'GQ', 'ER', 'ET', 'GA','GM','GH',
'GN','GW','KE','LS','LR','MG','MW','ML','MU', 'MZ', 'NA', 'NE', 'NG', 'CD', 'RW', 'ST', 'SN', 'SC', 'SL', 'SO', 'ZA', 
'SD', 'TZ','CF', 'CD', 'TG', 'UG', 'ZM','ZW')
AND ed.month = '2022-04'
AND mwh.snapshot = '2022-05'
AND mwh.event_timestamp >= '2022-04-01'
AND mwh.event_timestamp <= '2022-05-01'
AND ed.user_is_anonymous = FALSE 
AND mwh.event_user_is_anonymous = FALSE
GROUP BY 
mwh.event_user_id,
mwh.wiki_db
"

In [50]:

first_time_editors_april <-  wmfdata::query_hive(query)

Don't forget to authenticate with Kerberos using kinit

In [52]:

first_time_editors_total_april <- first_time_editors_april %>%
    filter(edit_count_group == 'first_time_editor')  %>%
    summarise(n_users = n_distinct(user_id))

first_time_editors_total_april

A data.frame: 1 × 1
n_users
<int>
3072

What language Wikipedias the people who are contributing from sub-Saharan Africa and who have made fewer than 100 cumulative edits are most active with?¶

Note: We'll use May to get a more recent depication.

In [60]:

first_time_editors_byproject <- first_time_editors_may %>%
    filter(edit_count_group %in% c('first_time_editor', 'under 100'))  %>%
    group_by(project)  %>%
    summarise(n_users = n_distinct(user_id)) %>%
    mutate(pct_users = paste0(round(prop.table(n_users) * 100, 2), "%")) %>%
    arrange(desc(n_users))
            
head(first_time_editors_byproject, 15)

`summarise()` ungrouping output (override with `.groups` argument)

A tibble: 15 × 3
project	n_users	pct_users
<chr>	<int>	<chr>
enwiki	3017	68.17%
frwiki	711	16.06%
ptwiki	143	3.23%
arwiki	97	2.19%
swwiki	67	1.51%
igwiki	41	0.93%
hawiki	37	0.84%
simplewiki	32	0.72%
yowiki	22	0.5%
dewiki	21	0.47%
rwwiki	21	0.47%
afwiki	18	0.41%
sowiki	14	0.32%
twwiki	13	0.29%
eswiki	12	0.27%

In [ ]: