Mining the richness of digital chat transcripts – An SMU Libraries experience in connecting with our community

• Main Authors: MAIDEEN, Shameem Nilofar, KOH, William
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2021
• Subjects: covid-19; text mining; chat service; digital transformation; chat transcripts; Library and Information Science
• Online Access: https://ink.library.smu.edu.sg/library_research/182; https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=1186&context=library_research
• Institution: Singapore Management University

The Covid-19 pandemic and the pivoting of library services to digital platforms have provided libraries worldwide the opportunity to extend the reach of their services to their communities. Where physical locations and operating hours used to be a barrier, the digital realm has freed many users to reach out and take advantage of the variety of services that are now truly available anytime, anywhere. This is being evidenced in the surge in access to electronic resources, the increasing numbers of attendees to online workshops, the upward trend in the use of chat services by the community, to name a few. In many ways, the pandemic has acted as an accelerant in the way our community has responded and adapted to those services and has had a positive effect on the impact, relevance and visibility of the library and its services. So, is there a downside? At SMU Libraries, digital transformation was a journey we embarked on intentionally in 2012, and this is reflected in our collection footprint and our move towards designing services that are digital-first. Key to this transformation was community input, and that was derived from the face-to-face interactions and the qualitative feedback that arose from those conversations. By removing opportunities for such interactions, we at SMU Libraries felt that we are missing out on rich data that has always helped us in creating effective services that truly meets the needs of the user. We saw an opportunity to address this gap in the increasing usage of our digital chat services. Our community was telling us about their challenges and giving us insights into their needs through these digital interactions. Moving beyond numbers and traditional quantitative metrics for measuring chat services, we decided to leverage on machine learning and AI, to draw out sentiments, to detect patterns and analyze user needs by mining and visualizing chat transcripts. This presentation will outline the proof-of-concept that we embarked on to visualize the chat services data, the insights and experiences we gained as part of this process, and how those will be used to better connect with, and contribute to, our community.