This year’s meeting of the Music Library Association was held in Salt Lake City (I attended virtually) in February.

Music and Wellness

A colleague at Baylor knew that music could, on many levels, address wellness needs, and was looking for a way to integrate this approach into her outreach. When she stumbled upon an early-1800s square (parlor-sized) piano in a storage room, she knew she had her answer. Victorian-style sing-alongs, with attendees gathered round the square piano and singing from sheet music from the archives (led by a couple of professional musicians), supplemented with related handicraft activities such as making your own calling card, were carefully designed around well-being principles (object-based learning, which incorporates slow-paced, sensory experiences, with scaffolding and learner-led interpretation) and trauma-informed care (somatic learning, based on sensation, movement, and breath, and providing safety in a low-pressure, rules-free environment). A series of these participatory musical soirees attracted 100+ students and faculty at each session, clearly filling a need across multiple campus communities, and garnering support for the music collection.

Student assistant training

Colleagues at North Texas wanted to boost student assistants’ confidence in their service-desk skills. They used a combination of the ADDIE model (Analysis, Design, Development, Implementation, and Evaluation) and the TILT model (Transparency in Learning and Teaching) to build a new training course in Canvas that includes online and in-person sessions. TILT emphasizes clear communication of purpose (course objectives, skills the students will learn, how success will be measured), grounding in real-life contexts (such as quizzes based on patron scenarios), and scaffolding (for common challenges in research assistance). Student self-ratings improved in long-standing areas of anxiety (taking phone calls, operating accessibility equipment, reading call numbers). Course evaluation by staff prompted decisions to make greater use of subject specialists, add content about the online catalog and special collections, and more frequent “touchpoints” or micro-assessments to “see what’s sticking.”

Music and AI

Colleagues have been busy developing AI tools to streamline cataloging of special collections. For a Harvard collection of 18th and 19th-century sheet music, a computer-science student was recruited to develop a program that uses Python and ChatGPT to search the publishers directory in IMSLP (a popular database of public-domain scores), and supplies an estimated publication date or date range. The Harvard folks hope to expand this application to more publisher bibliographies and catalogs, and possibly refine the dating based on additional factors such as prices in newspaper ads, or stylistic patterns in the cover art.

A similar approach was employed by colleagues at Stanford with a collection of Reggae 45rpm records. The challenges here included very scanty information on the discs, and artist stage names that are often colorful but brief, duplicate other names, or shift over time. The Stanford folks used AI tools to create a script that harvests data from discogs.com (a popular collectors’ site), then queries Wikidata for related identifiers that can be used to establish name authorities.