Designing a Data Visualisation for Interdisciplinary Scientists. How to Transparently Convey Data Frictions?

Abstract

This study investigates how the frictions that emerge while synthesising disparate datasets can be transparently conveyed in a single data visualisation. We encountered this need while being embedded in an academic consortium of four epistemologically-distant scientific teams, who wanted to develop new interdisciplinary hypotheses from their merged datasets. By inviting these scientists to collaboratively develop visualisation prototypes of their data within their own and then towards the other disciplines, we uncovered four data frictions that relate to discipline-specific interpretations of data, methodological approaches, ways of handling data uncertainties, as well as the large differences in dataset scale and granularity. We then recognised how the resulting visualisation prototypes contained several promising techniques that addressed these frictions transparently, such as retaining their overall visualisation context and using visual translators to mediate between differing scales. Driven by critical data discourse that calls for frictions to be foregrounded rather than be occluded, we generalised these techniques into a series of actionable design considerations. While originating from a single case of an interdisciplinary collaboration, we believe that our findings form a crucial step towards enabling a more transparent and accountable interdisciplinary data visualisation practice.