Analyzing and Visualizing Uncertain Knowledge: Introducing the PROVIDEDH Open Science Platform

Underlying uncertainty in DH research data affects decision-making and persists during the project's lifecycle. This uncertainty will always be present. Thus, efforts in providing technical support for humanistic research should focus on managing and making it more transparent, rather than removing it. Locating and tracing (certain types of) uncertainty through the evolution of a textual corpus can be done with the use of TEI tags. However, the use of these methods is not a common practice. The motivation of this paper is to address one possible barrier to wider use of these tags by providing a user-friendly interface to collaboratively annotating texts with uncertainty. We propose some minor extensions of the TEI specification that follow from our metrics of uncertainty. The first extension is adding the new “category” attribute to the “certainty” element, required to indicate the source of uncertainty. The second extension is to change the closed list of values of the “locus” attribute to an open list, in order to be able to explicitly indicate the attribute to which uncertainty refers. Additionally, the authors detected a need to describe the nature and type of uncertainties as well as evaluating the degree of uncertainty the piece of data introduces. Our tools on the platform were developed against the background of human-centered design with the focus onto easing uncertainty annotation and visualization, promoting the use of TEI standards and making uncertainty play a more active role in the research process