Recent Commons UW uploads by batch size

Last run on Dec 13 2017

Uploads within the last 30 days using Upload Wizard, grouped by user and minute. Grouping by user/minute is a proxy for "batch uploads." We assume that any files that were uploaded within the same UTC minute, by the same user, were uploaded in the same Upload Wizard batch. Likewise, we assume that images that were uploaded in different UTC minutes were parts of different batches. This approach will probably slightly under-count the size of batches, especially large batches, and slightly over-counts the number of batch uploads, because the larger the batch, the more likely it will overlap multiple minutes of UTC time. Nevertheless, this should give us a rough estimate of the total number of batch uploads, the relative proportion of batch-vs-single-file uploads, and the average size of batch uploads.

For reference, about 22,251 results were 'batch uploads' (more than one file uploaded per user-minute), and 27,708 were single-file uploads.

SQL

use commonswiki_p;

select rc_user_text, upload_minute, count(upload_minute) as batch_size 
from 
(select left(rc_timestamp, 10) as upload_minute, rc_user_text 
     from recentchanges where rc_log_type = "upload" 
     and rc_log_action = "upload" 
     and rc_comment = "User created page with UploadWizard"
     and rc_timestamp > "201711130000"
     order by upload_minute, rc_user_text)
as tmp group by tmp.rc_user_text, tmp.upload_minute order by upload_minute asc, rc_user_text asc;

SQL

Checking query status...