Information Visualization
Visualization is widely used to represent, view, and analyze data in the digital humanities. Standard line graphs, pie charts, bar graphs, and heatmaps for 2D matrix data are among the well-known visualization techniques for representing numerical data. The digital humanities also employ newer information visualization approaches to visualize non-numeric, qualitative data. The widespread adoption of these methods, most of which have been developed outside the humanities, however, contain problematic aspects. In a 2011 article, Johanna Drucker, Professor in the Department of Information Studies at the University of California, Los Angeles, criticizes the increased adoption of computational methods by humanities scholars, particularly visualization approaches that have been developed in other disciplines primarily for scientific use. She notes that maps generated by geographic information systems (GIS), and types of graphs intended for statistical display “conceals their epistemological biases under a guise of familiarity” (Drucker, 2011). She even refers to information visualization tools as a type of “intellectual Trojan horse(Drucker, 2011). The main problem, as Drucker sees it, is that what is presented on these plots are thought to self-evidently constitute “unquestioned representations of ‘what is’” (Drucker, 2011). She draws a sharp distinction between observer-independent “data”, which is “given” – that is, observable and recordable, and represented as (sometimes misleading) symbols on graphs, plots, pie charts, etc., and “capta”, which is actively “taken”, or constructed, and not simply a natural expression of some fact that exists independently of the observer. The observer-independence that characterizes data is not generally (or in a wholesale manner) consonant with observer-codependence of the phenomenon being investigated. The latter denotes investigation and activity that privileges the subjective and partial character of knowledge production, which is fundamental in humanities scholarship. Drucker warns that observer-independent visualizations, tools, and techniques developed by and for those in other disciplines conflict with the humanistic method, as the disciplines in which they were developed do not share the same “fundamental epistemological assumptions” (Drucker, 2011). She also states that when “the methods grounded in empirical sciences are put at the service of the social sciences or humanities in a crudely reductive manner, basic principles of critical thought are violated, or at the very least, put too far to the side” (Drucker, 2011). Drucker explains that an observation of a phenomena – in which empirical “data” are created for statistical representation and analysis – is not identical to the phenomena which is observed. In the case of common information visualization approaches, the critical distance is removed between phenomena and their interpretation, and thereby undermining the interpretative stance which underlies the production of humanistic knowledge.
Along with her criticisms, Drucker offers suggestions on integrating humanistic values with information visualization. She proposes a subjective model, applicable for humanities scholarship, in which treats data as capta. The model addresses four levels of interpretation or knowledge production:
- experience: modelling phenomenological experience;
- relationships: modelling and representing the relationships between documents and fields of discourse (for example, dates on diplomatic documents during World War II need to understood differently than documents from later years that refer to those dates);
- spatio-temporality: notions of space and time need to be modelled in documents used in the humanities, such as narratives; and
- performative quality: the interpretation of any or all the preceding characteristics also needs to be modelled.
Drucker offers an example using the well-known visualization of Dr. John Snow, who created a chart that led to the discovery of the Broad Street Pump as the source of a cholera outbreak in 1854. She proposes enhancing the original dot symbols used on the original with layers that express the individuality and lives of the persons affected, such as “profile, age, size, health, economic potential, family and social roles”, as an indication these features, representing an individual life, are not identical to or reductive to each other (Drucker, 2011). She concludes with the observation: “We have a very long way to go in creating graphical expressions that serve humanistic interpretation” (Drucker, 2011).
