Research:Editor Behaviour Analysis & Graphs
Duration: 2015-06 – ??
Contact: Jonathan Morgan
This page documents a research project in progress.
Information may be incomplete and change as the project progresses.
Please contact the project lead before formally citing or reusing results from this page.
The project aims to explore different way to visualise the entire edit activity of a wiki from its birth. The graphs should help us understand macro/community level changes in editor activity.
The graphs are broadly divided into two. The graphs with the data for articles and the same with the data for editors.
The wiki db's on toollabs are queried for the data & then transformed into graphs.
- Active Editor
- An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
- Editor Group/Cohort
- Editors are grouped by the month of their first edit.
- Active Article
- Similar to the active editor an active article is defined as an article which receives 5 or more edits in any month.
- Article Group/Cohort
- Articles are grouped by the month of their creation.
The high-level research questions that we want to answer visually are the following:
- What is the average longevity of editors and how has it changed over the years & across different languages.
- How does the contribution of new editors in their first month compare with existing editors?
- How is it in terms of edit sessions?
- How is it in terms of content/bytes added?
- How has it changed over the years?
- Are users finding it difficult to discover articles where they can meaningfully contribute?
- Is the reducing/plateauing editor count a result of a maturing encyclopedia?
- What does edit activity look like from an article's perspective?
- Turn the graphs into a dashboard to understand macro changes in a language/community wiki.
- As a dashboard for the grants & evaluation team.
- Longevity of editors on en
- It started falling in Jan 06 and has been stable since Jan 07. Since Jan 07 on average only 5% of the editors joining in a month were active after 4-6 months. Pre Jan 06 more than 5% were active for atleast 60 months. (https://cosmiclattes.github.io/wikigraphs/data/en/editor_cohort_longevity.html, select 5% and above in the filter. The first image in the gallery is a screenshot of the same.)
- New Editors vs Existing Editors on en
- Editor activity peaked in Jan 07 - Apr 07 (Image 2 in the gallery, https://cosmiclattes.github.io/wikigraphs/data/en/Editor_Cohort_Contribution_in_a_month_by_value_stacked_view.html). But the percentage contribution of new editors (editors who joined that month) in that month has remained the same (Image 3 in the gallery, https://cosmiclattes.github.io/wikigraphs/data/en/Editor%20Cohort%20Contribution%20in%20a%20month%20by%20percentage%20-%20stacked%20view.html). The fall since Jan 07 - Apr 07 was due to fall in new editors & also the older editors. Filter the graphs to by selecting 1-2 in the selector.
- The cohort that contributes most in every month is the cohort the joined in that month - the new editors.
- Longevity of articles
- In all of the languages which were analysed (ru, es, it etc) it is the articles that were created in the initial days (birth of the wiki to 2004 - 2005) that are continuing to get edited. The articles created since then get a storm of edit activity when they are created but subsequently see very little edit activity on them.
- Edit Activity on articles in a given month.
- Two major chunks of edit activity can be observed. the biggest % of edit activity happens in the articles that are created in the month in question. The second chunk of activity happens with the articles created in the early days of the wiki.
- The monthly decline in active editors in en may be attributed to the old timers. We see this across other languages too.
- Only the articles in the (beginning - 2007) cohorts continue to see active edit activity.
- Retention rates are much higher in languages like de.
- zh (continues to show uptrend on many fronts. Active Editors, retention etc).