As the Audiences department prepares to work on major improves to mobile editing, many teams are working to understand the high-level patterns of mobile editing (the success rate of edit attempts, the retention of new mobile editors, the revert rate for mobile edits, editors' proportion of mobile edits, and so on).
However, we'd also like to develop a deeper understanding of the content of mobile edits. Questions we'd like to answer include:
- What types of edits do mobile editors generally make? What's the breakdown between copyedits, factual changes to existing content, and additions to new content?
- How big are mobile edits? What's the size distribution among different edit types (e.g. copyediting one paragraph versus copyediting an entire 20 paragraph article)? How many sections of a page do mobile edits generally change?
- How often do mobile editors write on talk pages? How often do they receive responses?
- What's the quality of mobile edits? Overall, is it higher or lower than desktop edits? If it differs, what are the main reasons for that difference? For example, is it that edits contain more factual errors or is it that they're just more poorly formatted?
- What can we learn about users' motivations by studying their edit patterns? For example, can we segment mobile editors by types, size, and pattern of edits made?
- Do the results of any of the above differ significantly by key characteristics such as:
- Editor used (mobile wikitext editor, mobile visual editor, iOS app editor...)
- Geographic location
- Project
Note that these questions are meant to give a general sense of the area we like to understand better: obviously, it won't be possible to answer all of these questions (both because of time and data limitations), and there may be other similar questions that are worth answering. Much of this project is about exploring the data and identifying the most promising avenues for investigation.