For three years, I have been monitoring awards announcements, checking in on Jennifer LaGarde and Donalyn Miller’s starred-reviews tracking, and reading book review journals. Two years ago, I spent some time at the GAP School library organizing their collection and seeing what flew off the shelf and what didn’t. I began checking best seller lists, browsing bookstore displays, and paying attention to JMRL Top Checkouts blog posts.
I research to expand my understanding of reading trends and prominent authors in children’s literature. I expand my knowledge to make better reading recommendations to individual readers.
I have been brainstorming about other data sources I could consult. I decided to request a year’s worth of circulation data from JMRL. My goal was to take a closer look at which children’s titles and authors are being checked out.
Meredith Dickens, the library’s Collections Manager, sent me the FY25 circulation data. She included two files: one for physically circulating titles/items and one for digital downloads (Overdraft).
The csv file for physically circulating items included:
- Book title field. This often included a subtitle. It sometimes included a type, such as “periodical”.
- Author
- Number of holdings (i.e. the number of copies the library has)
- Checkouts (i.e. the number of times a title was checked out)
The csv for digital downloads (Overdrive) included:
- Title ID. I believe this is system-generated by Overdrive
- Title
- Checkouts (i.e. the number of times a title was checked out)
Looking at these files, I saw there was a wealth of information. As sent to me, the files did not include information about what was and wasn’t children’s literature or even a way to immediately filter out non-book items. To meet my goal, I would need to connect the information in the files to other data sources.
I considered asking Dickens if she had the necessary data. I realized that I also had an opportunity to embark on a side project: Learning how to use an AI Agent to code.
Method
I set out to use the AI Agent Claude to help me clean up the files and add information via OpenLibrary’s API with GoogleBooks API as a fallback option:
- I cleaned up the title field by separating it into two additional columns: subtitle and type. I used Claude to fix errors and edge cases.
- For the digital file, I used a Python script (written by Claude) to call OpenLibrary’s API and Google Books API to identify which titles were likely to be magazines. This introduced some errors that I could not resolve. For example: Is Vanity Fair the book or the magazine?
- I used Python scripts (written by Claude) to retrieve ISBN for book titles via OpenLibrary’s API and Google Books. For the physically circulating items, I had both title and author. This likely led to more accurate results for the titles where ISBN could be located. It also led to a higher “not found” rate because both title and author had to match. For the digitally circulating items, I only had title. The python script looked for ISBNs that corresponded to English language edition of the ebook. It used “most recent published” as a tiebreaker. This file has an 18% error rate. Some common errors: ambiguous cases where a title refers to two different books, fetching a related ISBN such as a study guide or derivative work, and just plain coding errors!
- Once I had the ISBN, I used it to retrieve the subjects associated with a title via OpenLibrary’s API and Google Books (python script written by Claude).
- I directed Claude to use the subjects column to populate a column called “Age-Range” with the words “Kid Lit” for each book that had a subject that indicated that it was children’s literature (such as “juvenile”). For the physically circulating items, I also asked Claude to take a second pass and identify books that were “Possibly Kid Lit” by considering the title and author fields combined.
- I took a second pass when I realized that Young Adult was being lumped in with Kid Lit and asked Claude to differentiate between the two by looking for Young Adult in the subject field
Results of File Cleanup
| Physically Checked Out Only | Digitally Checked Out Only | Combined Digital and Physical Checked Out | |
| Number of items | 124,688 (72% of total checkouts) | 48,249 (28% of total checkouts) | 172,937 |
| Percentage of books missing ISBN | 27,138 out of 111,378 Book titles (24.37%) | 8,292 out of 47,848 Book titles (17.33%) | 35,430 out of 159,226 Book titles (22.25%) |
| Percentage of books missing subject | 41,386 out of 111,378 Book titles (37.16%) | 16,060 out of 47,848 Book titles (33.56%) | 57,446 out of 159,226 Book titles (36.08%) |
Note: Dickens also told me that if a title had been removed from circulation, it did not appear in the original files Dickens sent to me.
Future Improvements
I would like to try improving ISBN matching through adding criteria to the Physically Checked Out file — favoring English language, print edition format, and publication date. I believe this would reduce ambiguity.
I would like to add “Author” to the Digitally Checked Out file. I could either use this to find more possible KidLit books or simply to analyze which authors are popular in the digital format.
I learned many things about coding with Claude that I will document in another post.
Initial Analysis
There were 1,400,607 checkouts of 172,937 titles in FY25.
All Items/Titles
Number of Items/Titles
There were 124,688 physical titles; 48,249 digital titles; and 172,937 combined physical and digital titles checked out in FY25.
| Physically Checked Out Only (% of physical titles) | Digitally Checked Out Only (% of digital titles) | Combined Digital and Physical Checked Out (% of Combined digital and physical titles) | |
| How many books were checked out that are identified as Kid Lit? | 35,260 (31.66% of Book-only titles) | 6,595 (13.78% of Book-only titles) | 41,855 (26.29% of Book-only titles) |
| How many books were checked out that are identified as YA? | 1,308 (1.17% of Book-only titles) | 928 (1.94% of Book-only titles) | 2,236 (1.40% of Book-only titles) |
| How many non-books were checked out? | 13,310 (10.67%) | 401 (0.83%) | 13,711 (7.93%) |
| How many items were checked out only once? | 32,465 (26.04%) | 9,163 (18.99%) | 41,628 (24.07%) |
| How many items were checked out in 5 times or less? | 84,261 (67.58%) | 26,427 (54.77%) | 110,688 (64.00%) |
| How many items were checked out 10 times or less? | 103,468 titles (82.98%) | 35,682 (73.95%) | 139,150 (80.46%) |
| How many items were checked out 50 times or more? | 1,255 (1.01%) | 1,545 (3.20%) | 2,800 (1.62%) |
| How many items were checked out 100 times or more? | 179 (0.14%) | 525 (1.09%) | 704 (0.41%) |
| How many items were checked out 200 times or more? | 20 (0.02%) | 172 (0.36%) | 192 (0.11%) |
| How many items were checked out 300 times or more? | 6 (0.00%, rounds down) | 83 (0.17%) | 89 (0.05%) |
Number of Checkouts
This table documents the number of check-outs for titles during FY25. There were 1,400,607 checkouts in FY25: 834,021 physical checkouts and 566,586 digital checkouts
| Checkouts of physical items (% of total checkouts of physical items) | Checkouts of digital times (% of total checkouts of digital items) | Checkouts of combined physical and digital items (% of total checkouts of physical + digital items) | |
| How many checkouts were there for titles with the type “book”? | 772,584 (92.63%) | 494,188 (87.22%) | 1,266,772 (90.44%) |
| How many checkouts were there for books labeled Kid Lit or Possibly Kid Lit? | 335,696 (40.25%) | 65,161 (11.50%) | 400,857 (28.62%) |
| How many checkouts were there for books labeled Young Adult? | 9,758 (1.17%) | 14,533 (2.57%) | 24,291 (1.73%) |
| How many checkouts were there for non-books types? | 61,437 (7.37%) | 72,398 (12.78%) | 133,835 (9.56%) |
Mean, Median, Mode of Number of Checkouts
| Physical Items – Checkouts | Digital Items – Checkouts | Combined physical + digital – Checkouts | |
| What is the mean, median, and mode of number of checkouts for books? | Mean: 6.94 Median: 3 Mode: 1 | Mean: 10.33 Median: 5 Mode: 1 | Mean: 7.96 Median: 4 Mode: 1 |
| What are the mean, median, and mode for books labeled Kid Lit or Possibly Kid Lit? | Mean: 9.26 Median: 5 Mode: 1 | Mean: 9.88 Median: 5 Mode: 1 | Mean: 9.36 Median: 5 Mode: 1 |
| What is the mean, median, and mode for the number of checkouts for books labeled Young Adult? | Mean: 7.45 Median: 3 Mode: 1 | Mean: 15.66 Median: 7 Mode: 1 | Mean: 10.86 Median: 4 Mode: 1 |
| What is the mean, median, and mode of the number of checkouts for non-book types? | Mean: 4.62 Median: 2 Mode: 1 | Mean: 180.54 Median: 23 Mode: 1 | Mean: 9.76 Median: 2 Mode: 1 |
Non-books Only
| Type | Number of Titles of Each Type that Were Checked Out | % of total number of Non-Book Items | Number Checkouts | % of Total Checkouts |
|---|---|---|---|---|
| Digital Magazine | 401 | 3% | 72,398 | 54% |
| DVD | 6,529 | 48% | 36,288 | 27% |
| Sound Recording | 6,105 | 45% | 16,576 | 12% |
| Read-Along Book | 300 | 2% | 3,662 | 3% |
| Periodical (Physical Magazine) | 180 | 1% | 3,221 | 2% |
| Kit | 61 | 0% | 969 | 1% |
| Blu-ray | 53 | 0% | 429 | 0% |
| Book Club Kit | 52 | 0% | 128 | 0% |
| Braille | 6 | 0% | 80 | 0% |
| Videorecording | 14 | 0% | 66 | 0% |
| Board Book | 1 | 0% | 5 | 0% |
| Electronic Resource | 2 | 0% | 4 | 0% |
| Audio-enabled Book | 1 | 0% | 3 | 0% |
| Large Print | 3 | 0% | 3 | 0% |
| Microform | 1 | 0% | 1 | 0% |
| Cartographic Material | 1 | 0% | 1 | 0% |
| Downloadable eBook | 1 | 0% | 1 | 0% |
| Total | 13,711 titles | 133,835 checkouts |
.
Discussion
There were 1,400,607 checkouts of 172,937 in FY25 from the Jefferson Madison Regional Library (JMRL).
That’s 3,837 checkouts a day. That’s 474 titles checked out per day.
Let’s imagine in. If each checkout is a person, that’s 3,837 of us walking into a branch or opening up a JMRL library app each day of the year. Four hundred and seventy-four titles begin circulating each day. How can there be fewer titles circulating than checkouts? Well, that’s because our librarians have stocked multiple copies of our most popular titles.
We read books and magazines on screens, but we still use our library branches. We’ve checked out over two times more physical titles as digital titles.
Whether we cracked open a book or fired up a screen, the majority of titles we checked out were checked out ten times or less (70-80% of the titles checked out). A majority of the titles were checked out five times or less (55% digital, 68% physical). We’re not all reading the same books, nor should we be. A healthy collection serves a wide range of interests and not just the best sellers.
Our children’s collection is important. Children’s books were 40% of the checkouts of physical items. When children’s titles are checked out, they are checked out over and over again (9 times on average).
Our digital YA collection is important. When a digital YA title was checked out in FY25, it was checked out an average of 15 times!
For non-book items, Digital Magazines are a powerhouse. Checked out, on average, 180 times. They account for over half of all the checkouts of non-book items. DVDs are the most popular non-book physical item. They account for about 24% of the non-book item checkouts.
I’m not a statistician. My files are flawed! Still, I think it’s safe to say: Our library is thriving.
I will continue to analyze the data. Next, I hope to dig into Kid’s Lit within the files.
Rachel Vogus' Habits and Hobbies 