Date(s) - 06/08/2016
2:30 pm - 3:30 pm
3910 15th Ave NE
Theme: Transparency, seamful design, and data science
We will have two short talks from Sarah Fox in HCDE (with demos from Meredith Lampe) and Jeff Heer in CSE. Then we’ll open it up for discussion around key questions in this area. What are the ethical considerations around transparency as well as how might a value of transparency be operationalized in data science practice, whether it is within algorithmic interfaces or data visualization? How visible or legible or “seamful” (as opposed to seamless) should particular logics and computational processes be within these interfaces and visualizations?
Sarah Fox’s presentation, “Digital interventions on Cartographic Visibility: Reflecting on and circulating socio-spatial differences through digital systems” approaches these questions of transparency and visibility of digital systems from a human-centered design perspective. She will discuss the recent visualization of the holes of service in the Amazon prime case: Amazon doesn’t consider the race of their customers, should it? among others. Also, Meredith Lampe will present demonstrations of their recent mapping experiments
Jeff Heer will discuss some of his work putting into practice some of these ideas in visualizing the results of machine learning methods. His talk will draw from this paper on the topic and discuss considerations of trust and interpretability.
Further reading if interested:
Algorithms and experiments make strange bedfellows at SXSW. Wood, 2016, ComputerWorld
A path to understanding the effectiveness of algorithm awareness, Hamilton et al. 2014, alt: CHI