2017 Charleston Library Conference

This presentation will focus on a use case in which information from library scientists helped data scientists craft a methodology for better understanding cyber risk behaviors. You will hear the first-hand story of the experiences and lessons learned in building the relationship between library and data sciences.

The project referenced in this use case was a data experiment based on statistical inference, the process of drawing conclusions about populations or scientific truths from data. Although there are many theories and modes for analyzing inference, this project focused on the complexities of the performance of the solution and the correctness of inference. In other words, how good could we get at educated guessing for missing, bias or unobserved data, and how well could we perform as a team to provide a proposed solution for different problem sets to the same enterprise problem – cyber risk.

In data science, it is about reliability and repeatability for any solution to be trusted, in library science it is about resources and provenance. How could we define, design, and provide references of measurable data attributes for components and features of a taxonomy for the data science problem? Our working hypothesis became the thought for an interactive intelligent taxonomy for the data scientists to reference as an aid to build a better data driven equation/computation/algorithm.

From the use case story, you will be provided background on the data science process, how library science intersects in that process, how data gets messy very quickly, and how rules can go out the window. Perspectives are a key component to a good working relationship between data scientists and library scientists. Get to know and understand the differences in the communities that will help you build a better data project.