The IDEA Institute has broadened my knowledge of artificial intelligence and deepened my understanding of machine learning in the academic library settings. In addition to the two months onboarding preparation when we learned programming language using Phyton, the weeklong intensive institute has force me to look at data curation from brand-new perspectives. It challenged my assumptions and checked my biases in every aspect of data mining and lead me to specifically think about these issues:
• Who collects the data?
• What kind of data to collect?
• Why is the data collected?
• When is the data collected?
• Where is data originated from?
• Who owns the collected data?
• Where is the data stored?
• How long will the data live?
In preparation to implement the Bento-style search result, the outcome depends heavily on the integrity of data in our existing library guide system platform provided by Springshare. We will have to adopt a consistent approach to label the subject headings for each library guides, preferably using Library of Congress Subject Headings. We also need to standardize our selection of keywords for tagging each library guides. The week-long boot-camp provided the cohort of librarians many opportunities to collaborate with colleagues from other institutions with different expertise. The metadata librarians in my interest group helped me realized the scope of my project may be over-ambitious and recommended to break it up into two phrases: To clean up the metadata in all tag entries, before we can adopt the Bento-style search results when we migrate to LibGuide 2.0.
While at the Institute, we also learn about how to draft the SMARTIE goals for our project design and implementation. My SMART goals are intended to be Specific, Measurable, Achievable, Relevant and Time-Bound, however it neglected elements of Inclusivity and Equity. The metadata librarians in my interest group reminded me that image alt-texts are also one form of data that is discoverable by search engines, such as google image. In order to our vision-impaired patrons to access image contents, we should also ensure all image titles are entered into the alt-text fields before our migration to LibGuide 2.0, in order to improve search accuracy and increase result relevancy.
