As digital humanities projects bring together more data than ever before, ways of presenting and accessing this data are transforming rapidly: text mining becomes more powerful and ‘searching’ and ‘browsing’ become more complex and interrelated. These tools not only help us find what we are looking for; they also shape our paths of inquiry. For this reason, it is important to carefully consider both the ontologies behind the tools (what exists and how can it be ordered?) as well as their epistemological effects (how do classification and representation affect our understanding?). How can we preserve, in a digital environment, the spirit of serendipity and discovery that characterizes humanities research? By asking such critical questions, we hope to forge tools that may generate unexpected results and new perspectives.
This paper addresses such considerations in the selection and design of navigation tools for humanities subject portals, as well as some current options and concerns related to these choices. As a concrete case study, we reflect on our experiences creating the Environment & Society Portal (http://www.environmentandsociety.org), a not-for-profit education, research, and outreach project at the Rachel Carson Center for Environment and Society in Munich, funded by the German Federal Ministry for Education and Research.
The Environment & Society Portal, which launched in early 2012, makes digital multimedia and interpretive materials in environmental humanities freely and openly accessible to academic communities and the interested public internationally. To enhance user engagement, in the coming year we will be building on features that facilitate and encourage feedback, interaction, contribution and dissemination of information.
The Portal’s multimedia content mixes retrodigitized environmental humanities materials with born-digital content, such as interpretive exhibitions, short descriptions of places and events, or the localized environmental histories related to broader issues found in the sub-project ‘Arcadia: European Environmental Histories’ (this collaboration with the European Society for Environmental History eventually plans to include all world regions). Content may be linked or clustered according to diverse research paths. It will be possible to add new entries on the same topics to provide a fuller picture or alternative perspectives.
The Portal employs three custom-designed interactive navigation tools (map viewer, timeline, and keyword explorer) that invite users to explore content of interest and to visualize chronological, spatial, and conceptual connections. While the Portal offers full-text indexing and searching as well, these three interconnected navigation tools represent its unique functionality. Similar tools exist elsewhere, of course. But rather than simply apply existing tools, we have reviewed our assumptions about categorization and representation and have either modified existing taxonomies or created our own controlled vocabulary; we have reviewed multiple benchmarks, considered the purpose of our representations, and have created our tools’ designs from scratch. While we have done user testing and can discuss our Google Analytics data, our main objective is to reflect on how data selection, categorization, and representation create constraints and opportunities for inquiry. We offer an examination of our decisions regarding the creation of these three navigation tools, and reflect critically, six months after our launch, on their performance.
Navigation via the Portal’s map viewer is possible by zooming and browsing on a map and/or searching for placenames. After considering several alternatives, we chose to use GeoNames’ highly hierarchical, but open-source, gazetteer (adopted also by other major European digital humanities portals such as Europeana). We took this decision to reflect as accurately as possible the complexities and vagaries of geographical and administrative dependencies, to potentially ease the retrieval of nested data (e.g., showing results related to a certain village among the results for its parent country), and to allow for a greater degree of interoperability with other websites. The most difficult and important representational consideration was our decision to represent only point data, not lines or areas. This was not only a pragmatic technical decision, but serves the purpose of our map viewer as a content index, not as a thematic mapping of a coherent data set.
The Portal’s timeline feature allows users to scroll forward and backward in time, to zoom in and out, and to plot the results of up to three searches on an interactive timeline canvas. One of the most difficult decisions was limiting the temporal metadata to years as points in time, to the exclusion of periods, eras, and specific or ‘fuzzy’ dates. Representational considerations included how to distinguish between search results for the same year and the use of a graphic scale for time. Furthermore it was important to us to present the Timeline as a content index, not as a ‘comprehensive’ collection of events. The decision to represent the quantity of relevant Portal results in the timeline slider/scale bar as connected ‘peaks’ may be one of the Portal’s most problematic representations.
While the map and timeline allow users to compare results geographically and temporally, the keyword search tool offers thematic navigation functionally resembling the former Google ‘wonder wheel.’ Specifically, our keyword explorer lets users refine a search by choosing among the tags that occur most frequently in relation to a given keyword, using a narrowing search and, in future, also a more exploratory, expansive approach in which keywords are linked without limiting the results to previously identified keywords. To facilitate this, we chose to adopt – and constantly develop – a flat controlled vocabulary instead of imposing a predetermined hierarchy of themes and keywords. This means that search experiences may change substantially as new materials are added to the Portal.
As new content is added, users’ search possibilities will grow exponentially. That is, it will be possible to use any tool to refine searches made with any of the other tools. The aim is to allow the users to create completely personalized paths through environmental scholarship and materials gathered on the Portal.
As the ways we organize and represent information have important consequences for the communication and use of knowledge, an enhanced navigational platform provides opportunities to rethink our usual approaches. We hope these tools will offer multiple ways to find and compare information, inspire alternative historical frames, and encourage the formation of new connections across disciplinary and political boundaries.
By examining the complex ontologies behind navigation tools and the epistemological considerations concerning their application and use, we hope our experiences and critical reflections may offer new perspectives for our colleagues working with digital humanities data and databases.