PetScan is a really handy tool for finding unreferenced articles on various topics. Here is a short guide to using it successfully.
The simplest way to start is to enter the category you want to search within into the Categories box under the Categories tab, and run the query. For us, this is Articles lacking sources. Don't include the namespace, that is, Category:, just the name of the category Articles lacking sources. This petscan does just that - click on the link to load it up. You'll see that this actually returns no results, and you might wonder why! This is because Articles lacking sources doesn't contain any articles, only sub-categories which themselves contain articles. To recurse down one level, and get all those articles, you need to append |1 to the category name thus: Articles lacking sources|1. Now, when you run this petscan, you'll get all the tens of thousands unreferenced articles. Remember this recursion technique - it is important for more complex queries.
Now, you probably want to filter this big list down and select only unreferenced articles on a particular topic. There are three main ways to do that - filtering by category, filtering by WikiProject and filtering by predicted topic or predicted country.
This is probably the simplest way, but slightly more prone to false positives, and it can be slower to return results. Add the Books category below Articles lacking sources in the Categories box, and be sure to choose the Combination → Intersection radio button. Again, Books is a category that contains mostly subcategories, and it takes a bit of recursion to get to actual articles. So choose a sensible recursion level - maybe Books|4 - and run the query. Do experiment with recursion levels here - the bigger the number, the more articles you'll get back, but that comes with a greater chance of returning articles that aren't actually books, because of the spider web that is Wikipedia's category system.
This is a bit more complicated, but not much, and it tends to throw up fewer false positives and return results more quickly. Instead of doing the intersection of two categories, carry out a query that selects only unreferenced articles whose talk pages have been tagged with {{WikiProject Books}}. Make sure that Categories box under the Categories tab has our base category and recursion depth Articles lacking sources|1 in it, and then select the Templates&links tab, and enter WikiProject Books into the Has any of these templates box, being sure to check the Use talk pages instead check box underneath it. Then run the query.
All articles are classified using machine learning and associated with both predicted topics and predicted countries. We can use these with Petscan to filter unreferenced articles too. Start with a clean query. Then, navigate to the Other sources tab, and enter into the Search box something like articletopic:books deepcat:"Articles lacking sources" for articles whose predicted topic is Books, or articlecountry:ind deepcat:"Articles lacking sources" for articles whose predicted country is India. You can replace the highlighted articletopic value with any of those listed at mw:Help:CirrusSearch/articletopic, or the highlighted articlecountry three-letter code with any of those listed at mw:Help:CirrusSearch/articlecountry. Then enter enwiki into the next box and choose a maximum number of search hits to return, and run the query.
PetScan has a load of options, and it's worth spending fifteen minutes or so having a play. Some things to try
| Articles beginning ABCDEFGHI |
(pagepile) |
| Articles beginning JKLMNOPQR |
(pagepile) |
| Articles beginning STUVWXYZ1234567890 |
(pagepile) |
| Articles, no lists, beginning VWXYZABCDEFGHIJKL |
(pagepile) |
| Articles, no lists, beginning MNOPQRSTU1234567890 |
(pagepile) |
| All probably unreferenced articles |
(pagepile) |
If you run into difficulties, feel free to get in touch with any of the editors listed below.