AI and the Sporadic Librarian-Coder: Adventures in Muddling Through and Excelling (Occasionally)

For the occasional librarian-coder--those who know more than the basics, but for whom coding is not a primary/daily task--trying to remember specific syntax or commands, like how to format a visualization in ggplot or how to execute a semi-complicated filter in a pandas dataframe, often creates enough inertia to prevent us from tackling data-related projects where we might add value in the library. AI coding tools, like ChatGPT and GitHub’s Copilot, can lower the bar enough to make such projects worthwhile and successful...at least occasionally. This presentation will showcase how I've used these tools to augment my less-than-professional coding skills in devising solutions to practical work challenges, like exploring and parsing new datasets, collecting data for patrons, making data purchased from vendors more usable, and improving the functionality of crappy database interfaces that I'm forced to tolerate. In addition to highlighting specific project examples, I'll discuss opportunities, challenges, and some things I've learned along the way about what can increase the chances for success as well as when to consider throwing in the towel.