Review: “Vis Week” in Berlin

In late October, a broad international research community dedicated to visualising information and data gathered in Berlin. In a great concurrence of events (termed “Vis Week” by some people), the grand annual IEEE vis conference was accompanied by the smaller interdisciplinary conference Information+ in Potsdam.

IEEE vis is THE most important platform for visualisation research rooted in the computer sciences, usually hosted in Northern America. This year’s installment in Berlin gathered some 1200 attendees from all over the world.
As a first time attendee, the conference to me seemed part competitive exhibition, part inspiration fair, and part global class reunion (hence also the informal party track happening every night alongside the day-time events). With three main tracks (Information Visualization, Scientific Visualization and Visual Analytics), with an impressive number of interesting speakers, papers, posters and satellite sessions, the conference is impossible to survey in total, but let me convey a few snapshots.

Graphic recording of Dieter Schmalstieg’s IEEE vis keynote speach. © Benjamin Felis

The keynote speech by Dieter Schmalstieg (Graz University of Technology) discussed using AR-technologies for visualising information and data in situ. Schmalstieg argued that while notions of Augmented Reality have been around for a long time, the technology has not been applied to visualisation tasks, due mostly to a lack of display technologies that would facilitate seamless interaction, as well as persisting difficulties in the technical workflow (i.e. mapping the virtual spatial model to the real world). Yet he made a case that augmented reality could make for very effective visualisations, for instance when users demand an overlay of data or information when looking at real world phenomena.

While the SciVis track covered visualisations of complex spatial phenomena or mathematical models, the InfoVis track focussed more on visualising numerical and statistical data. An ongoing theme at the conference is the quest to support the comprehension of multi-dimensional data, for instance by providing new tools for exploration, by including automated annotations or creating new designs for complex visualisations. Another theme is the creation of collaborative workspaces, where teams can interact with a visualisation across various devices, such as Vistrates. A fascinating new move is the attempt to automate the “design expertise” which has evolved in the field (i.e. rules of “how to” or “what not to”) in visualisation tools. Two papers presented different approaches (Draco from the University of Washington and Litvis from City University London). Alper Sarikaya et al. provided an illuminating survey study of the mundane genre of dashboards, claiming that this type of visualization—albeit ubiquitous in everyday life—does not receive enough attention from the vis research community.

“What is Bayesian Knowledge Tracing” (Young Cho et al.) – a project presented at the Visualization for AI Explainability Workshop

One very complex theme wafted through various sessions under the umbrella term “explainable AI”—i.e. the search for appropriate visualisation tools to leverage an understanding of the inner workings of complex AI models. This could be helpful to experts for controlling the results of their model, as well as for communicating to non-experts what the model is doing (“building trust” was a keyword here). The workshop on Visualization for AI Explainability explicitly called for ideas and approaches, for instance to help investigate training data sets or to create visual representations of the decision cycles of a given AI model.

“Artificial Senses” was one of the works presented at the VISAP exhibition. The project creates an artistic visualisation of sensor data gathered while using smart phone. © Kim Albrecht/Metalab Harvard

The IEEEvis conference was accompanied by the VIS Art Program which featured a selection of data driven art in an exhibition. To my perception, the exhibition—titled “Data and Identities”—was a successful contribution in that the featured projects added artistic perspectives to the technically and mathematically oriented mind-set that traditionally dominates the IEEE vis.

On the weekend before IEEE vis, the University of Applied Sciences in Potsdam hosted the second installment of the interdisciplinary conference Information+, which attracted a more diverse crowd from the fields of design, geography, journalism and the humanities. I had the honor to open the conference with a keynote about historical information visualisation, in which I argued that studying historical examples helps us set up a critical mind-set in judging and creating data visualisation.

The project “Simulated Dendrochronology” by Pedro Cruz et al. was presented both at Information+ and VIS Art Program. It shows immigration to the US as rings of a tree – beautiful, precise and intuitive.

This notion was mirrored by several other talks which also provided critical perspectives on current habits of data collection and data visualisation. Among these speakers were Catherine d’Ignazio, who—from a feminist perspective—investigates the notion of “objectivity” and shows how statistical data sets tend to unveil the discriminating biases of their creators. Her argument can be retraced in her recently released manuscript “Data Feminism”. Similarly, Ron Morrison discussed in their closing keynote the unintended consequences of quantifying and visualising data on minority groups and or those with less power in society.

For everyone slightly interested in the visualisation of data, this week provided plenty of food for thought (too much to really digest, actually). The talks of Information+ were recorded and should be up on the conference website sometime soon. For more input regarding various IEEE vis conference papers, I recommend the blog eagereyes written by Robert Kosara or the Datastories episode 130.

——-


Sandra Rendgen is the author of two bestselling books on data visualisation and infographics, both published by Taschen. Currently she is preparing a book on the work of Charles-Joseph Minard, one of the most important forefathers of modern information visualisation. It will be published by Princeton Architectural Press on November 6, 2018.

Share