User:SuggestBot/Documentation/Suggestion columns

From Wikipedia, the free encyclopedia
Jump to: navigation, search

This page attempts to explain what information is displayed in each of the columns in the table SuggestBot posts to users who receive suggestions periodically. If you find it confusing or have questions, please do not hesitate to let us know on SuggestBot's talk page, we'll do our best to help you out and improve this documentation.

Example suggestion table and explanations[edit]

Picture with an excerpt of a set of suggestions from SuggestBot on English Wikipedia, showing the full set of columns available.
Picture with an excerpt of a set of suggestions from SuggestBot on English Wikipedia, showing the full set of columns available.

There are nine columns in the table with SuggestBot's suggestions. Note that the table is sortable, meaning that it can be sorted by any of the nine columns. If you want to have the most-viewed articles towards the top you'll click twice on the Views/Day column and the table will end up sorted in descending order by that column. It is also sometimes possible to sort by two columns, for instance if you first sort by popularity (as mentioned above) and then sort by one of the task columns, you'll get the most popular articles needing a specific tasks towards the top of the table.

We will explain the columns one by one in order from left to right, except for the five tasks columns which will be explained together.

Views/Day[edit]

The number shown in this column is the average number of views per day each article had over the two weeks prior to the suggestions getting posted, so it's possible to gauge an article's popularity. We use the number of article views as reported by the volunteer run site Wikipedia article traffic statistics. A two-week average is used to smooth out daily variations while also allowing it to be influenced by sudden spikes in activity.

Quality[edit]

This column shows a prediction of an article's quality on a Low/Medium/High scale to allow for quick assessment. We have used a sophisticated piece of machine learning software (called a Random forest) to provide SuggestBot with some understanding of Wikipedia article quality, which it then uses to predict which of the seven assessment classes a given article falls into. The seven classes are then collapsed into Low/Medium/High quality as follows: Stub- or Start-class means Low quality, B- or C-class means Medium quality, while A-class, Good Article, or Featured Article means High quality.

Holding the cursor over an article's quality prediction brings up more information
Holding the cursor over an article's quality prediction brings up more information

SuggestBot also checks if an article has been assessed by a WikiProject. This information is partly used to override its predictions, if an article has been assessed as A-class, Good Article, or Featured Article it will always be marked as High quality. Secondly, the article assessment is displayed upon moving the cursor over a given article's quality rating. The image on the right shows an example of this, the article Dorian mode is predicted to fall into the Good Article class, giving it a High quality rating (three stars), while it is currently assessed as Start-class.

For those interested in a scientific discussion of how we predict article quality, see our article published in the proceedings of WikiSym 2013: Tell Me More: An Actionable Quality Model for Wikipedia

Title[edit]

The title of the suggested article as a link to the article itself. For convenience we also provide a link to the article's talk page.

Task columns (Content/Headings/Images/Links/Sources)[edit]

Each of these columns will feature a red X if SuggestBot has calculated that an article could need a specific type of improvement. There are five such columns, each aims to answer a specific question:

Content
Is more content needed?
Headings
Does this article have an appropriate section structure?
Images
Is the number of illustrative images about right?
Links
Does this article link to enough other Wikipedia articles?
Sources
For its length, is there an appropriate number of citations to sources in this article?

SuggestBot uses Wikipedia's featured articles as the gold standard of article quality and only adds the red X if an article appears to be severely lacking in quality for a specific task compared to the Featured articles. If a red X is shown there is also a popup available by holding the cursor over the X that provides a short summary of the suggested task, e.g. "Please add more content" for the Content column.

Keep in mind that all columns in the table are sortable, thus if you want to focus on a specific task you can easily find all suggested articles matching that task by sorting on that column. Say you want to work on finding sources, clicking on the "Sources" header will sort the articles by that column, bringing all the articles that appear to need more sources to the top of the list

Tagged with…[edit]

This column shows which of SuggestBot's task categories the article was picked from. SuggestBot only suggests articles that the community has identified needs improvement, e.g. "Add sources", and "Cleanup". For a more thorough explanation of which task categories SuggestBot has and which Wikipedia categories it grabs articles from, see User:SuggestBot/Documentation/Task categories.