Research talk:AfC processes and productivity/Work log/2014-04-17

From Meta, a Wikimedia project coordination wiki

Thursday, April 17th[edit]

Today, I replicated the last analysis, but I trimmed out edits made by the creating user so that we can see activity by other users.

The geometric number of revisions per week by non-creating users is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
Revisions by others. The geometric number of revisions per week by non-creating users is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
The geometric number of bytes changed per week by non-creating users is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
Bytes changed by others. The geometric number of bytes changed per week by non-creating users is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
The geometric number of unique, non-creating users per week is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
Unique non-creating users. The geometric number of unique, non-creating users per week is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
The geometric number of unique, non-creating anons per week is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.
Unique non-creating anons. The geometric number of unique, non-creating anons per week is plotted for AfC drafts and Direct to Main articles that were deleted quickly or not.

If your article is going to survive the first couple of weeks, then you're uniformly better off starting it in the main namespace. If it's not going to survive, then AfC may provide a higher amount of interaction with and contributions by other users, but only since AfC switched into full gear. The relationship between AfC and Direct to Main article creations that were deleted quickly is likely to be dominated by the effect of immediate deletions of many Direct to Main articles that fail to assert notability. --EpochFail (talk) 12:51, 17 April 2014 (UTC)[reply]


I think I'll quickly replicate these plots without splitting the deleted/not-deleted. --EpochFail (talk) 12:51, 17 April 2014 (UTC)[reply]

OK. I'll save the bits. It's the same story as yesterday. In other words, removing the creating editor from the equation didn't have a meaningful effect on the interpretation of the results. However I think splitting the articles by those that were deleted quickly (within hours) and those that were not has given us insight. Next I want to look at the size of the creating revision. --EpochFail (talk) 13:05, 17 April 2014 (UTC)[reply]

The empirical PDF of the number of bytes saved in the first revision to newcomer created articles is plotted by origin, the AfC Epoch, and whether the article was deleted quickly.
Creation size distribution. The empirical PDF of the number of bytes saved in the first revision to newcomer created articles is plotted by origin, the AfC Epoch, and whether the article was deleted quickly.

OK. Fun story. If you're going to create an article that's going to get deleted. It's going to be about 1k in size. If you create an article directly in main that's about 50 bytes, then it is very likely to survive (tiny stubs that are obviously useful?)

If you create an article in AfC, it's very unlikely to be shorter than a 200 bytes. --EpochFail (talk) 13:23, 17 April 2014 (UTC)[reply]


I just kicked off another aggregation process to filter out revisions to AfC drafts that change the status so that we can filter out some (most?) of the process stuff. --EpochFail (talk) 13:26, 17 April 2014 (UTC)[reply]