#DPLWhatsNext: How Dallas Public Library Engages Readers Using Linked Data
By Kate Meador, Pauline Stacchini, and Andrew Wright at Dallas Public Library
Dallas Public Library (DPL) is a 29-location urban library system that serves seven million customers annually. With the support of a NoveList “Engaging Readers at Your Public Library” grant, we established the social media campaign #DPLWhatsNext to promote our readers’ advisory resources to the public.
This campaign is a Personal Librarian program we launched in February 2018. Nine library staff serve as personal librarians who curate monthly reading/watching/listening lists based on their diverse interests. These lists — ranging from “woke picture books” and “brooding sci-fi” to “Texas poet laureates” — provide multiple areas for staff to connect with customers’ interests.
Each personal librarian also has a profile on the #DPLWhatsNext website with a brief synopsis of their subject focuses, such as “horror movies,” “true crime,” or “LGBT.” Customers can browse these profiles, choose a personal librarian, and ask for suggested titles in either their area of interest or an unfamiliar genre. Customers request their personalized recommendations by filling out a brief online form describing their preferred genre, age range, and format. Those who don’t want to select a specific personal librarian have a “Surprise me!” option.
We brought linked, structured data into the #DPLWhatsNext program to solve our greatest challenge in building the site: displaying the personal librarians’ lists.
We needed a way to group multiple bib records into a carousel, and to display multiple carousels unique not only to each personal librarian but also to each new monthly list. And we needed statistical usage data on all of it.
Our solution was to add a local subject heading, exclusive to each monthly list’s topic, into every bib record of every list. To achieve this, the personal librarians used Polaris’s record set feature to construct their monthly lists. Cataloging then determined the list’s local subject heading and bulk changed it into the bibs.
The Library.Link Network support team created a sandbox we could upload bibs into to test the process prior to launch. After several trials, we chose the following subject heading format:
650 7 ‡a PLP ‡x Gossip ‡v Biography ‡y 2018.01. ‡2 local
Once the records were processed by the Library.Link Network, the local subject headings became unique Concepts in DPL’s local graph of linked data. Each list could now be retrieved via its associated Concept, and the content carousel embedded on our #DPLWhatsNext page. The carousels of linked data let us track statistics for each item.
After going live, we fine-tuned the process after two major discoveries:
1) Content carousels that contain an open entry bib will show items for every ISBN on the record. Our first manga carousel had 100+ items for 15 titles, creating display issues on some browsers.
2) The controlled language of our original local subject headings didn’t improve customers’ user experience. We now use uncontrolled vocabulary to increase user relevance and aid searchability in our structured data:
650 7 ‡a PLP ‡x Gardening ‡y 2018.04. ‡2 local650 7 ‡a PLP ‡x Fantastical beasts fiction ‡y 2018.04. ‡2 local
We only have one month’s statistics at the time of this writing, but in that month the content carousels were viewed 728 times and led to 2,748 catalog clicks. And we are working with the Library.Link Network to refine these numbers by distinguishing between the humans and bots that access our data (without collecting personal identifying information about users, of course).
With time and continued promotion, we predict these numbers will increase.
By connecting multiple tools — NoveList, staff expertise, and the Library.Link Network — DPL launched a new readers’ advisory initiative with both seamless customer experience and significant statistical data. The Library.Link Network was a key part of this, not only because of its content carousels’ ready-to-embed coding, but also for its reporting ability on what titles receive more clicks and which lists engage customers most.
That usage data allows us to move forward with smarter choices when selecting personal librarians and their topics, while also tracking the highest-circulating material types. So, what’s next for #DPLWhatsNext? We’re currently working on a project with our local PBS broadcaster to create content carousels using an adaption of the same process.