Do Reference Librarians Dream About Complicated Reference Questions?

I woke up dreaming that one of my colleagues in Edmundston was requesting a shelflist report to assist in building a regular rotating 'Library Bestseller' display in his library. (Which would probably make a lot more sense if he were the Adult/Young Adult Services Librarian, but that's dream-logic for you, isn't it?)

My brain being the rogue tool that it is, it has of course been pondering over how to build such a report (even though no one really asked for it, outside of a dream) instead of grabbing another hour's sleep.

So what parameters would you want to automagically identify library bestsellers suitable for a display? (Given sufficient ops-time, I suspect it's probably possible for a sufficiently experienced human to build such a display by grazing the shelves and pulling books based on their cover recognition. But sometimes it's fun to see if you can figure out criteria that the computer can apply to hack the same problem.)

The obvious criteria would, I suppose, include:

• Available on the shelf at 'my' library (Library to be definable at the point of scheduling the report)
• Total checkouts (at the title-level, not the copy-level) more than X (X to be definable at the point of scheduling the report)
• Optional criterion to limit the report to books published since year Y (Y to be definable at the point of scheduling the report)
• Optional criteria to limit the report to particular item types (fiction, non-fiction, etc.), item category 1s (large print, CD audiobooks, etc.), item category 2s (adult, young adult, etc.), and/or statistics category (donations, Aboriginal collection, etc.)

Would one want the ability to limit the report to a particular genre? It'd probably be possible to hack on a seltext query to search the 099 field in the bibliographic record (similar to Gaby's Book Club Report*) and pipe it through to the item-level filters.

And then there would be the necessity/desire to limit the report to titles that are still somewhat fresh, even if they are no longer hot enough to have holds to fill. After building the list of candidate titles based on the other criteria, we could calculate the last date on which any of the copies were checked out, sort the list by that date, and head the first 25 (40? 50? How many books do you need to build a display?) to get the most recently checked out (and presumably still interesting) titles.

On the output side, if you are going to use this to pull a display from the shelves, you'd want it sorted by home location, then call number, right? And then the usual suspects for the other data? Author, title, publication year, item barcode. This would probably be a report that one would want scheduled and the output emailed, so a columnar display rather than tab-delimited (for dumping into Excel, etc.)

So, that's mapped out, so my brain can stop chewing on the problem! Fortunately, my rogue brain does not require me to actually write the Perl scripts to get these things out of the back of my head, but it'll sometimes keep poking me with an interesting question until I at least map out how I could tackle the challenge :-)

(* Yes, I have a bad habit of naming custom reports in Symphony after the people who first asked for them. Most of the present and past TSLs, ARDs, and Circulation Supervisors have their names stored in source code on the server for posterity, as do a smattering of library managers and other staff. We send her mental well-wishes every time we schedule irene to suspend holds newly placed for pickup at a library that is closed for renovations.)