Visualizations for the Queries Deleted pages , Automated Edits , Edits revert,rollback,undo in Wikimedia from Multiple wikipedia language sets (tewiki , hiwiki , mlwiki) #8
Conversation
Number_of_Automated_Edits.sql
Outdated
| JOIN user_groups | ||
| ON actor.actor_user = user_groups.ug_user -- Join user_groups and actor tables | ||
| WHERE user_groups.ug_group = 'bot' -- Filter for bot user group | ||
| AND revision.rev_timestamp BETWEEN '20230101' AND '20240301'; -- Filter by specific date range |
There was a problem hiding this comment.
don't filter in WHERE condition, that way you won't be able to filter on the dashboard.
Add DATE to SELECT statement. Also, count distinct revision IDs.
| SELECT | ||
| (SELECT COUNT(*) | ||
| FROM revision -- For edits we use revision table | ||
| WHERE LEFT(rev_timestamp, 8) BETWEEN '20240101' AND '20240301') AS total_edits, -- Filtered between specific dates |
There was a problem hiding this comment.
We only need deleted pages, not revisions. This can be removed.
|
|
||
| (SELECT COUNT(*) | ||
| FROM archive -- For deleted pages we use archive tale | ||
| WHERE LEFT(ar_timestamp, 8) BETWEEN '20240101' AND '20240301') AS deleted_pages; -- Filtered between specific dates No newline at end of file |
There was a problem hiding this comment.
Similar to above, don't filter in WHERE condition. Add date statement to SELECT
| SELECT | ||
| (SELECT COUNT(*) | ||
| FROM archive -- For deleted edits we use archieve table | ||
| WHERE LEFT(ar_timestamp, 8) BETWEEN '20230301' AND '20231212') AS deleted_edits, -- Filtering between specific dates |
There was a problem hiding this comment.
deleted edits are not required, only pages, this can be removed.
| (SELECT COUNT(*) | ||
| FROM revision r -- For reverted or rollback edits we use comment table | ||
| JOIN comment c ON r.rev_comment_id = c.comment_id | ||
| WHERE LEFT(r.rev_timestamp, 8) BETWEEN '20230301' AND '20231212' -- Filtering between specific dates |
There was a problem hiding this comment.
remove filtering in WHERE on timestamp, add them to SELECT statement as date
| AND (c.comment_text LIKE '%revert%' | ||
| OR c.comment_text LIKE '%rollback%' | ||
| OR c.comment_text LIKE '%undid%')) AS reverted_edits; -- Comparing strings i.e revert or rollback with column comment_text No newline at end of file |
There was a problem hiding this comment.
This is not a good way to get reverted edits. You have to join the ctd_tag and ctd_tag_def tables, and check if the tags are mw-reverted, mw-rollback or mw-undo.
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
yeswanth120-gif
changed the title
| @@ -0,0 +1,464 @@ | |||
| { | |||
There was a problem hiding this comment.
This graph looks very clumsy. For representing plots with continuous data, line plots look more cleaner instead of bar plots. With a line plot, this graph will have just 3 lines each with a color and it also records the ups and downs over time without looking clumsy. It becomes easy to understand insights
Reply via ReviewNB
| @@ -0,0 +1,464 @@ | |||
| { | |||
There was a problem hiding this comment.
Line #22. if all_dfs:
Box plots are not the ideal way of representing this. Histograms, bar plots (including count plots), and line plots are effective for visualizing data where "number of counts" is a key metric. Histograms are ideal for displaying the distribution of numerical data, while bar plots, especially count plots, are suitable for showing the frequency of categorical variables.
Reply via ReviewNB
|
Please review my updated Visualizations notebook in reviewnb |
yeswanth120-gif
changed the title
3 participants
Metric 1 : Number of automated edits
Metric 2 : Number of deleted pages and edits
Metric 3 : Number of edits deleted, reverted or rolled back