Visualising data: Research by Mereke van Garderen

Visualising data: Research by Mereke van Garderen

By Mereke van Garderen

Information visualization is defined as the study of visual representations of abstract data to reinforce human cognition. In simpler terms, what I aim do is to make pictures of data in such a way that it becomes easier for people to interpret the information. This is a concept that many people are familiar with. For example, we all learned to draw a pie chart to look at proportions back in high school. What many people do not know is that visualization is an actual science. A good visualization is not just determined by personal taste and creativity. Although it certainly helps to have some of both, there is a lot more to it. Visualizing archaeological data is particularly challenging for multiple reasons. We often encounter multivariate data, a mix of qualitative and quantitative information, a temporal and a geographical component, biased sampling, missing information, etcetera. Furthermore, the diversity in data types and objectives within the NEXUS1492 project is huge. People study data ranging from ocean currents to survey observations to ceramics production and many other things. This calls for a wide variety of visualization methods as well.

The objective of visualisation is two-fold: on the one hand, visualization can help a researcher explore the data and make observations that would not be apparent by just looking at a data table. On the other hand, a visualization can be used to communicate results we already found to a larger audience. This is especially important for dissemination and public outreach, a key component of the NEXUS1492 project. My thesis consists of three parts; The first part is an overview of various different visualization methods that have applications in archaeology and heritage studies, many of which I have worked on in collaborations with other people from the project.
The second part of my thesis focuses on analytic methods that are well known in mathematics or computer science, but so far had not been applied in archaeological research.
The third part focuses on visualizing georeferenced data: showing information on a map. This poses particular challenges in the context of the Caribbean, where we have many small islands with densely clustered information, and large empty areas of ocean in between. There are two aspects to this problem: finding a good representation for the information about one data point, and placing these representations on the map such that the symbols do not overlap.

I designed multiple representations for different kinds of data that we encountered in this project in the form of glyphs, small symbols that represent multiple variables at once. One example, of a glyph representing a museum with Caribbean collection, is shown here. If we would place these on the map at the exact museum location, the symbols would overlap. Removing overlap is a widely studied visualization problem, but none of the existing methods was well-suited for the kind of heterogeneous, geographical data we were dealing with in this project. I therefore developed a new algorithm to layout symbols on a map without overlap. This method, which will be published in June, can help not only archaeologists but anyone dealing with geo-referenced data to communicate their data more effectively.

19-06-2017

Top