Why and how do we use visualisation in our work? 3rd Nedimah Infoviz workshop
Report from 3rd NeDiMAH infoviz workshop
Visual Tools and Methods in Digital Humanities: Capturing, Modelling, Reading, and Thinking about Knowledge Creation...
7-8th of March 2013 in Umeå, Sweden.
Why and how do we use visualisation in our work?
The workshop explored visual components in digital environments and interpretative research tools and queried the shifting intersection between descriptive and analytical uses of visualisations. We discussed 'new' readings both within and beyond traditional textual modes, and questioned shifts in representation within the digital sphere. This 'reading' is a part of the whole research process. However there is a need to increase the awareness and conceptual thinking about how we form our research with visual methods.
The workshop discussed critical reading principles both for research itself but also for creating digital tools for different aspects of the research process. The workshop was organized so participants had to prepare themselves by reading selected papers in advance, thereby making it easier to understand the different steps involved in a range of research methods and how different decisions affect the knowledge we aim to create. The workshop gathered 11 participants mostly from the NEDIMAH working group. There were also guests from other working groups of NEDIMAH as well as other invited expert participants.
In the workshop we shared a variety of perspectives on how to use, and how to reflect upon the use of visualisation in the humanities. The participants shared presentations and had prepared themselves by reading selected contributions and also by preparing comments and questions. This framework facilitated an active and participatory engagement from the beginning, as each person had read and thought about the others’ work prior to the meeting. Moreover, all the speakers gradually embedded the previous presentations in their talks, using them as examples or terms of comparisons.
This critical engagement was invaluable as it allowed participants to move beyond a surface description of their research and method, into comparative analyses from which shared points of concern and interest emerged.
While preparing and reading others’ contributions the participants tried to consider the different phases of the research process, so questions and comments could be considered in the wider context of information visualization.
Such research phases were identified as:
• the formulation of the research question,
• data capturing, modelling and representation,
• selection of data,
• modelling of data,
• modelling search, query conditions and possibilities
• visual representation for descriptive and analytical purposes,
• communication of findings in publication and for peer-reviewing.
Overall we aim at aligning the finding to the planned survey of NeDiMaH in order to make the result easier to communicate overall . Therefore the workshop started with an overview of the work with the Survey and how it could be used to collect knowledge that is valuable for our work and for the scholarly community.
The areas of research covered different areas of humanistic research, such as:
• Representing hidden elements in city landscapes for urban planning,
• Modelling and representation of spatial and temporal uncertainty
• The role of perception and the effect of thinking in 3D
• Visual methods of linking entities in networks with uncertain sources
• Visual methods of seeing patterns and entities in statistical data for example with spatial support
• Reflecting and debating upon aesthetics in dataviz
• Accessing and evaluation of interfaces for facetted browsing
Key Outcomes The papers covered a variety of topics and methodologies. On the one hand, they gave a panorama of the richness and complexity of the Digital Humanities, and they highlighted some common issues and the importance of establishing methodological standards.
One of the main topics discussed was how to deal with fuzzy and ambiguous data, especially from a visual point of view. The public, and sometimes even other scholars, tend to assume that there is a simple one to one relationship between information and its representation on a digital map or in a database. Things are seldom that simple, and levels of uncertainty should be clearly perceivable and understandable by the final users of digital tools. As some of the papers specifically pointed out, we need a new shared vocabulary; a kind of visual grammar that will, gradually, became a guideline.
All the papers showed how different digital humanities techniques can be used not only to visualise and communicate the outcome of other disciplines but also to discover new information and ask new questions. Digital tools are not always able to give answers to research questions if used on their own, but they are very effective in identifying gaps, inconsistencies and anomalies, pointing out what needs to be investigated further.
It is quite common that research focuses on multiple interpretations allowed by ambiguous archaeological finds. For this reason, it was useful to consider the stress put by one of the papers (Visualisation in History) on the scholars’ interpretative biases due to their own cultural backgrounds. However, the paper on narratives in World War II (Marten During), prompted deep thinking on the political agendas of each country in different time periods; agendas that are very easily detectable after some years, but less visible when the researchers themselves are immersed in it.
Another discussion that was very useful was on a user-centred approach, working mainly with and for the citizen. Digital tools and services that are not used will be dead in few years, in spite of how interesting and stimulating they could have potentially been. From a more practical point of view, underused digital tools will make it more and more difficult for future ones to be funded. Issues of sustainability and user engagement are critical in designing integrated shared solutions for research.
The project about ancient rock carving showed a very effective way to engage a young audience with cultural heritage, trying not only to make the topic as much appealing as possible but also finding a language that makes easier for the user to communicate with the interface and perform rewarding searches.
Moreover, digital tools (such as database, websites, 3D models in Real Time) should interact with the local communities, museums and cultural institutions (as advocated, for example, by the Portable Antiquities Scheme).
A number of key issues emerged from our conversation, all crucial to the participants that might be guidelines or principles for knowledge creation and design:
• The idea of the digital moving beyond what was previously possible, foregrounding multiplicity and multivalency instead of a more linear and two dimensional approach.
• The need to keep in mind personal / plural users ourselves and others… multiple engagements.
• Awareness of persistent themes and decisions “what to leave in and leave out” in terms of perception and data quality.
• A conviction that engagement with uncertainty was good – how much do you encourage users to engage? Recognize that user engagement allows them to fill the gaps, not consume a prepackaged story and being prepared to relinquish control of the narrative
• Enthusiasm about the amount of work using linked data.
• Recognition that we need to define the object of study/the domain while trying to define research questions and approaches to the object of study?
• Sensitivity to the delineation of boundaries is important.
• Discussion of motivation and what forces move us between these defined entities causing these motivations to move, research goals, learn.
• Connectivity: how we do model relationships between these defined entities and what kinds of networks of collaboration would be useful?
• Projects should be relevant to the domain that we are studying and to interpretation of the domain
• We need more possibilities to envision uncertainty, to use imagination to combine things together.
• Greater interdisciplinarity (Swedish genealogy example) would result in a place to reinvent the information and the visualization.
• Secret concept in tools – Johanna Drucker says that new tools are like Trojan horse, bringing their epistemological assumptions with them from their home disciplines.
• The autonomy of method of the researcher is very important and we want to be autonomous.
• Visualising ambiguity and uncertainty (fuzziness) is crucial, our example is the use of cartographic symbology to indicate and make ambiguity transparent.
Reusability of data and methods and generalizing our approaches to it allows us to think about and question the choices we make. What research questions are being supported by the environment. All were struck by the reusability of Fredrik's faceted browsing environment for the Byzantine Jewry project and Matthias' Genealogical explorations . All participants recognized the importance of documenting the building of visualisations.
Shared language around visualisation is emerging, such as thinking about During’s ' use of symbols to bridge cultural gaps while also introducing our own complexity.
In During’s example of modelling driven by the research question, multiple perspectives must be included rather than being distilled into a single point of view. There is a common language of modelling, (eg of the CIDOC CRM), but it is important to see multiple possible combinations and interpretations moving towards a Linked Open Data or semantically aware set of models or interpretations
We need too to consider how to use interactivity, and refine our process through engagement, capturing user interaction for the augmentation of a dataset and further refining it in the process. We suggest that key aspects of engagement are to involve the user and to help them re-member the knowledge being explored or represented, but also the central concept must be to negotiate for truth. The idea of public engagement and play was considered also as well the question “what can we do to make our visualisations easier to use?”
With regard to uncertainty, it seems useful to build a contrast between the seeming perfection of visualisations while using them to highlight the imperfection - and further the perceived authenticity of text, visualization and numbers. Returning to the theme of fuzziness and ambiguity, we discussed the question of making users aware that data are not perfect and how can we visualise this imperfect data. How can we use visualisation not just as an end, but as a tool and a methodology to challenge narratives and interpretations