Language:

ProgressiVis relies on a Python interpreter that is convenient for quick prototyping but is inefficient. Once the semantics of the language and its core libraries become clearer, we will replace the implementation with a C++ library that will improve performance and allow its integration in other popular data science environments such as R and Julia. We are collaborating with colleagues in the database community to tackle the challenges related to progressive data management  139, 62.

Algorithms:

Existing libraries implementing algorithms are either “eager”, sometimes “online”, and more rarely “streaming”, but very few are compatible with our progressive paradigm. We have started to work in collaboration with the University of Seoul and the University of Delft on the problem of converting existing algorithms into a progressive form. We currently have a running version of approximate k-nearest-neighbors  22, kernel density estimation  22, k-means clustering, multidimensional projection using t-SNE  22, and several other core algorithms. We want to further expand the range of algorithms we can support, and better understand how algorithms could be transformed into their progressive counterpart.

Progressive computation can trade speed, memory, and quality. When analysts want to compute rough estimates of their results quickly, several progressive algorithms offer parameters to control the tradeoffs. For example, our progressive k-nearest-neighbors algorithm  22 allows choosing the number of k-d-trees that influences the memory footprint, accuracy, and speed. Progressive data analysis also uses and adapt streaming algorithms, also called “data sketches”  89, that offer various tradeoffs between speed, memory, and quality.

Visualizations:

Like algorithms need to be adapted or transformed, visualization techniques and user interfaces also need to be adapted to be used in a progressive setting. For the visualization pipeline, including the rendering phase, there has been previous work related to progressive layout computation and progressive rendering. For example, in network visualization, there is a old tradition of using iterative force-directed solvers and showing their results progressively when computing complex graph layouts. We want to generalize this approach to all the visualizations. In the recent years, we have collaborated on work related to the progressive computation of multidimensional projections, including PCA and t-SNE  129, 22, and will continue to improve the scalability of projection methods. Progressive rendering is popular in the real-time computer graphics and gaming domains, but new in visualization. Our colleague Renaud Blanch from Grenoble has started to investigate this problem  73 by designing a rendering engine that can visualize and filter a large amount of data in interactive time, transiently simplifying details when necessary. We will also explore this problem both in the context of direct rendering, when the GPU shares the memory performing the computations, and web-based rendering.

We have also started to explore the problem of requirements for progressive user interfaces  61, leading the community of researchers to understand that progressive systems need to provide effective assessments of the quality of their results to allow analysts to make early decisions. This justifies the next research topic on the management of uncertainties. We have also started to propose new data models for managing aggregated visualizations  109. We will now continue our research to adapt existing visualization techniques and interaction environments to deal with progressive results and parameter steering.

Management of uncertainties

When performing exploratory data analysis on large data, analysts start by trying some algorithms with different parameters until they obtain useful results. Therefore, it is of paramount importance to be able to realize that a specific algorithm or a specific set of parameters are not leading to a useful result and to abort an analysis before it ends to explore other methods. On the other side, in many circumstances when data is “well-behaved”, progressive algorithms achieve very good results quickly, and analysts can make a decision before the end of the algorithms' execution. The assessment of a quality measure related to a computed result is a complex problem, already tackled by other communities (e.g., streaming, approximate queries). Progressive Data Analysis adds two problems: 1) controlling uncertainty during progressive computations, and 2) conveying the results of this uncertainty in a way understandable by analysts, while avoiding possible cognitive biases such as priming and anchoring  111.

To control the uncertainty during the progression of a computation, methods related to Sequential Analysis  90 should be used. They are well-known in the area of clinical trials and have evolved recently to manage large amounts of data while controlling time or quality in Approximate Query Databases such as BlinkDB  58. These approaches should be further adapted for progressive data analysis. The topic of human cognitive bias will be developed further in sec:decision.

3.2.1 Scientific Foundations for Data Physicalization

In the next few years, we will work on establishing and consolidating the new area of data physicalization. Data physicalization is an emerging cross-disciplinary research area that examines how computer-supported, interactive physical representations of data, that we call “data physicalizations”, can support cognition, communication, learning, problem solving, and decision-making  106, 44. In contrast to traditional graphical on-screen data visualizations, physicalizations show data in a material form.

Although the research community, and especially the members of Aviz, have made first steps by conducting initial experiments  104 and laying out a research agenda  106,1 the area still lacks a theoretical and empirical foundation that can guide the design of physical data representations, both static and dynamic  118, 119.

Data physicalizations can potentially offer a useful complement to flat (paper-based or screen-based) visualizations due to the ability of physical objects to tap on human skills such as active perception, depth perception, and non-visual senses; they may have the ability to support learning via direct manipulation, and can possibly engage broader audiences. However, there is currently little understanding of the underlying mechanisms, and the exact situations where physicality may be beneficial.

In particular, work in educational psychology and tangible user interfaces has shown that physical manipulation can facilitate cognition, while research on vicarious learning has suggested that people can also learn from observation. However, it is not clear which approach is the most effective, and no study so far has started to examine how previous findings generalize to data physicalizations. We will investigate to what extent manipulable physicalizations can support data-driven cognition compared to non-manipulable physical media such as paper, and to manipulable virtual media (e.g., computer graphics). We will investigate the perception, recollection and interpretation of a data physicalization when people watch it being animated vs. they manipulate it themselves, vs. they watch a third-person manipulate it. We will attempt to conduct this research in collaboration with Barbara Tversky (Professor Emerita of Psychology at Stanford University, Professor of Psychology and Education at Columbia University, and President of the Association for Psychological Science), with whom we have been in close contact.

We will also be working on a taxonomy of physicality, in order to better untangle the factors that distinguish physical from virtual displays, and to help the community develop a coherent terminology that will serve as a thinking and communication tool.

Although we have been involved in research on digital fabrication in the context of data physicalization  134 and despite the new and exciting advances in this domain, we might not pursue this line of research so we can better focus on the research questions mentioned previously. In other words, we will focus more on why build data physicalizations than on the how.

3.2.2 Situated Data Representations for Personal Analytics

In addition to studying physical representations of data, we will study how embedding data visualizations in the physical world can empower people to make effective use of their personal data in a variety of application contexts. People generate data in many areas of daily life (e. g., data generated from shopping cards, fitness and health data). However, these types of personal data are often only exploited by the companies capturing it while they remain unused by the people who “generated” them. Rarely is personal data subjected to an in-depth analysis and used for reflection and decision making.

Personal visual analytics is a new and growing area of research 93, but has mostly focused on desktop and mobile applications. With our work we want to go beyond the traditional platforms of personal data analytics by using situated data visualizations. In a situated data visualization, the data is directly visualized near the physical space, object, or person it originates from 145. For example, a person may attach small e-ink displays embedded with sensors at various locations of their house or their workplace, to better understand their use of space, of equipment, or of energy resources. Or a person who wishes to exercise more may use an augmented reality device to visualize their past running performance in-place. New situated data visualizations like these can surface information in the physical environment—allowing viewers to interpret data in-context and take action in response to it. Because they will make it possible to explore personal data and take action in relevant physical locations, situated visualizations of personal data have the potential to radically transform how people relate to data.

This research will investigate how situated visualizations of personal data can help address problems faced by the general public by investigating a range of research hypotheses and application domains. For example, we will study how augmented reality technology can help surface digital traces of personal data in physical environments and enable people to make more informed decisions. We will also study how small, portable, and situated micro-visualizations can allow people to understand personal data in-place in changing, dynamic, and mobile contexts.

On this topic, we already have an “équipe associée” Inria grant with University of Calgary, and we are currently submitting an ANR PCR proposal in collaboration with Sorbonne Université and Université de Bordeaux.

3.2.3 Visualization using Augmented Reality Devices

Based on past work on the integration of various types of input and output devices for the visualization of inherently three-dimensional spatial data 65, 63, 67, 97, 122, 100, 101, 102, we plan to investigate the use of, in particular, augmented reality (AR) head-mounted displays (HMDs). This combination of stereoscopically projected 3D data which provides a high amount of visual immersion with innovative interaction techniques that facilitate data exploration though immersion that arises from direct manipulation is part of the recently established subfield of visualization research which focuses on immersive analytics 86, 75 and promises to combine benefits of traditional VR-based environments with those that use traditional, touch, or tangible input and used innovative interaction designs. We specifically will focus on AR-based HMDs because they do not prevent us from using traditional data exploration environments such as workstations which are not only familiar to experts but also widely available. The AR HMD, moreover, is an easily maintainable device that does not rely on constant calibration and thus could more easily be integrated into and accepted by existing research facilities than more elaborate VR settings such as CAVEs.

Specifically, we currently have started two projects in this direction. One is a collaboration with particle physicists at the local linear accelerator laboratory as well as researchers at CERN with whom we are exploring an interactive data exploration tool of the experimental data generated at CERN from collisions of protons. In a second project we are working with climate researchers to explore novel collaborative data analysis scenarios in which several experts investigate 3D datasets together with the help of AR HMDs.

3.2.4 Other

In addition to the three major themes previously mentioned, Aviz has also been involved in research on the use of tangible input devices, both to navigate abstract data visualizations  105, 114 and to explore 3D data  101, 68, 66. We will continue this stream of research, possibly applying emerging trends in tangible input device design  91 to data exploration.

3.3.1 Perceiving the Gist of a Visualization

It would be highly inefficient if users of large data visualizations serially inspected individual visual elemnts. Instead, the visual system allows people to quickly get a visual gist of an overall pattern in the image  136. But how quickly and reliably can we detect these gists when multiple groups or regions in a visualization need to be summarized seperately and compared? What happens to speed and accuracy as more items are added to a visualization? These open questions about the impact of data scale and complexity on visualization perception were discussed in the review article by Szafir, Haroz, Gleicher, and Franconeri  136. To answer these questions, the experiment methods and findings from visual perception can be employed. Specifically, “ensemble coding” is the term used by vision scientists to describe our ability to rapidly perceive simple statistical properties of sets without the need to inspect individual items. Research has shown that we can perceive averages for sets made of simple  59 and complex  135 objects as well as the variance of sets for simple objects  127. But it is not clear if or how we can answer more complex statistical questions about subsets within visualizations, such as:

• How different are trends between groups of line charts?
• Which group's trends are more variable?
• Is the distribution of trends uniform or normally distributed?

For the field of visual perception research, understanding how we make these statistical decisions with collections would be a novel contribution that helps explain how our brain represents and summarizes complex information, how much information is discarded to make rapid judgements, and when biases or accuracy limits arise when working with collections of items.

For the field of data visualization research, understanding the conditions when we can make these statistical judgements and the speed and precision with which we can make them will directly translate to new guidelines for optimal use of visualization for large datasets. Although much visualization software uses settings that rely on design intuition, these results would be the first science-based guidelines for large data visualization and progressive visualization.

3.3.2 Cognitive Limitations for Visually Selecting Subsets in Visualizations

A critical task for analyzing large data mentioned in the previous section is filtering and comparing subsets. But why and when are these filtering tasks difficult? And how can a visualization facilitate the identification of subsets – especially for data with many categories?

How much information we can visually select and attend to and the minimum size of a selection are strongly limited, and this limitation is impacted by the layout of items in an image and the types of visual features and tasks employed  92. Past visual cognition research on the cause of attentional limits  94 generally presumed that we focus on information in a roughly circular region. But subsets in a visualization are often irregularly shaped, and the limitations of attention selection for more chaotic images like visualizations are not understood. Researching how people select irregularly shaped subsets would be a novel contribution to our understanding of visual cognition and could inform how progressive visualization techniques can show datasets such that all subsets can be visually compared.

3.3.3 Illustrative Visualization

We will continue our work in the area of illustrative visualization. This sub-field of visualization takes inspiration from illustrators' decades to centuries of experience in using knowledge from perception and cognition to better portray scientific subject matter. Another input arises from the field of non-photorealistic rendering which has developed numerous techniques of stylizing images and other input data. Traditionally, illustrative visualization has thus been applied primarily to data with a concrete spatial mapping in 2D and, more frequently, in 3D space. Nonetheless, some efforts for understanding the potential benefit of illustrative visualization to non-spatial data have been documented, including some of our own team 80, 60. An interesting aspect in this context is the role of depiction style can play in how visualizations or other forms of depiction are interpreted and understood. For example, we recently investigated the effect of stylization on emotion for images 69, and will extend this work in the future to study how we can use stylization to change the way people perceive, consume, and understand visualizations. One hypothesis to investigate, for example, is that a more hand-drawn depiction style 85 could lead to viewers spending more time with a visualization, and thus take more time to understand and interact with it.

Another main future direction of research in this context is the question of what the role of abstraction is in illustrative visualization 141 as well as visualization in general, and specifically how we can provide dedicated means to control the abstraction being applied to visual representations of data. This means that we need to go beyond seeing abstraction only as a side-product of stylization as it has traditionally been viewed in many approaches in non-photorealistic rendering as well as illustrative visualization to date, and investigate how we can interactively adjust it to provide practitioners with a means to find the best visual representations for a given task. For example, we have investigated this question recently in the context of structural biology 125, 146 and DNA nanostructures 124, 123, but also want to expand this work to other application domains in the future.

3.3.4 Decision Making with Visualizations

Aviz has recently started a novel stream of research on the topic of judgment and decision making with visualizations  78, 77, 79. Human decision making and cognitive biases are important research topics in the fields of psychology, economics and marketing, but until recently, the presence of cognitive biases in visualization-supported decision making has not been investigated. Yet, visualization systems are increasingly used to support decision making: large companies switch to visualization solutions to improve their decisions in a range of areas, where large sums of money or human lives are at stake. More and more, the ultimate goal of visualization is not to understand patterns in the data and get insights as was traditionally assumed, but to make good decisions. In order to fully understand how information visualizations can support decision making, it is important to go beyond traditional evaluations based on data understanding, and study how visualizations interact with human judgment, human heuristics, and cognitive biases.

We will pursue this important stream of research by investigating yet-unexplored areas, such as group decision-making with visualizations, decision making in the presence of uncertainty and incomplete information (in connection with the topics discussed in Sections 3.1 and 3.4.2), and the use of visualizations to support social choice and group decisions in the presence of conflicts of interest. How cognitive biases interact with visual perception is also an important and difficult question that has remained largely unexplored. In terms of application domains, we will be focusing more on evidence-based decision making in everyday life, e.g., how can visualizations be useful (or harmful) when choosing medicines, food products, means of transportation, charity donations, or when electing politicians. For example, a recent study has investigated whether specific infographic designs can lead people to feel more empathy towards suffering populations  74, but with inconclusive or negative results. In collaboration with the University of Campina Grande in Brazil, we are currently investigating designs that we believe are more likely to be effective. We will be evaluating these designs using metrics that are more likely to be correlated with prosocial decision making than empathy, as it has been shown that the emotion or empathy is not always conducive to helping behavior.

3.4.1 Promoting and Following Open Research Practices

Fundamental tenants of scientific research are that it is accessible, scrutinizable, and can be built-upon. For visualization research, the low rate of accessibility of both published articles and supporting materials makes it almost impossible to question the claims made in publications or to extend them. We measured the state of some of these issues for the IEEE VIS conference, finding serious problems across every facet of openness 115:

1. Open Artifacts: Questioning the veracity of research claims is difficult if the supporting evidence and methods are not available for scrutiny. Without these supporting research artifacts, reviewers and readers cannot thoroughly verify, replicate, or reproduce research, and readers may have difficulty extending and applying research. Such artifacts include but are not limited to: Code, resource files, parameters, sample datasets, and documentation needed to reproduce software, toolkits, or experiment procedures Raw empirical data collected or analyzed to support a paper’s conclusions Analysis code or scripts used to support a paper’s conclusions
2. Transparency in Decisions: To clarify distinctions between exploratory and confirmatory research, to reduce hidden flexibility in research (i.e., unreported researcher degrees of freedom  144, and to minimize publications bias, researchers must be educated about the impact of these issues and use approaches such as preregistration and registered reports where appropriate  128115.
3. Transparency in statistical reporting: Researchers must report the outcomes of their statistical analyses in a way that does not prompt misinterpretations. This aspect will be further developed in the next subsection.

For these issues to be addressed, educational materials and guidelines need to be written, so researchers have clarity about how to make their research more credible. Moreover, Aviz members are working with the organizing bodies of the visualization research community to establish incentives for making research artifacts and potentially establish minimal requirements for openness in published articles. Meanwhile, it is important to continue measuring and cataloging openness in the field to monitor progress. The goal is to improve the credibility and applicability of the field’s research.

3.4.2 The Communication of Statistics Results

Statistics are tools to help end users accomplish their task. In experimental sciences, to be qualified as usable, statistical tools should help scientists advance scientific knowledge by supporting and promoting the effective communication of research findings. Yet many areas of experimental science have adopted tools that have proven to be poor at supporting these tasks  81. One very common but particularly damaging practice is “dichotomous inference”, i.e., the classification of statistical evidence as either sufficient or insufficient, for example by looking at whether a p-value is above or below the .05 threshold  64. Dichotomous inference encourages practices that distort the scientific literature, such as publication bias, outcome reporting bias, and significance chasing  140. All of these issues contribute to making published studies less trustworthy and less likely to replicate.

A commonly advocated solution is to focus on effect sizes and their interval estimates rather than on p-values. However, interval estimates do not offer a sure protection against dichotomous inference  64. Therefore, many modern statisticians and methodologists urge researchers to think of evidence as gradual rather than binary when interpreting their results, irrespective of the statistical methods used. However, there is a lack of guidance on how to do to so. We will be conducting research on how to design statistical charts that prompt nuanced interpretations, and more generally, charts that convey statistical uncertainty in a honest and transparent manner. The two main targeted populations will be i) experimental scientists (who use research papers as a medium to communicate about scientific findings) and ii) the general population (who use newspapers and online articles to inform themselves about recent science advances). One goal will be to develop guidelines for faithful statistical communication that are evidence-based instead of being based on personal opinions. This research lies at the intersection of statistical cognition  70 and information visualization.

Besides studying visual representations that encourage non-dichotomous inferences, we will pursue one line of research we just started at Aviz, which considers the research paper as a user interface, and seeks to look at how this interface can be improved to help scientists communicate their results in a more transparent manner than using conventional research papers  84.

As a longer-term goal, we are planning to develop conceptual models in order to better understand and better reflect about the diverse phenomena underlying data-driven scientific communication. This includes the process of communicating empirical data between scientists, from scientists to people, and from journalists to people: we will attempt to better understand and describe the entire information pipeline, where information can be lost or distorted, and how to make the pipeline more robust to [self-]deception  83.

3.4.3 Shaping the Scientific Visualization Community

In the past, Aviz researchers have been heavily involved in the organization structure of IEEE visualization conferences, the most prestigious conference in our field, by proposing workshops, tutorials, serving on various organizing committees, steering committees, editorial boards, and even the IEEE Visualization Executive Committee that oversees the whole IEEE Visualization conference. We will continue to do so but will, in addition, be involved in newly created committees that help to restructure the community and prepare it for future growth and inclusivity. We will, in particular, aid the process by providing data collection and analysis services through the vispubdata.org dataset that we are collecting, cleaning, and making available to the community. The dataset has already been used in research (e. g., 117) but also to shape the scientific community by proposing program committee members, new processes, and is currently used by the Visualization Restructuring Committee (ReVISe). We are also involved in the EuroVis community and participate at multiple levels to its organization and management.

7.1.1 Cartolabe

• Name: Cartolabe
• Keyword: Information visualization
• Functional Description: The goal of Cartolabe is to build a visual map representing the scientific activity of an institution/university/domain from published articles and reports. Using the HAL Database, Cartolabe provides the user with a map of the thematics, authors and articles . ML techniques are used for dimensionality reduction, cluster and topics identification, visualisation techniques are used for a scalable 2D representation of the results.
• News of the Year: This year, Cartolabe was applied to the Grand Debat dataset (3M individual propositions from french Citizen, see https://cartolabe.fr/map/debat). The results were used to test both the scaling capabilities of Cartolabe and its flexibility to non-scientific and non-english corpuses. We also Added sub-map capabilities to display the result of a year/lab/word filtering as an online generated heatmap with only the filtered points to facilitate the exploration.
• URL: http://www.cartolabe.fr/
• Contact: Philippe Caillou
• Participants: Philippe Caillou, Jean-Daniel Fekete, Jonas Renault, Anne-Catherine Letournel
• Partners: LRI - Laboratoire de Recherche en Informatique, CNRS

7.1.2 BitConduite

• Name: BitConduite Bitcoin explorer
• Keywords: Data visualization, Clustering, Financial analysis, Cryptocurrency
• Functional Description: BitConduite is a web-based visual tool that allows for a high level explorative analysis of the Bitcoin blockchain. It offers a data transformation back end that gives us an entity-based access to the blockchain data and a visualization front end that supports a novel high-level view on transactions over time. In particular, it facilitates the exploration of activity through filtering and clustering interactions. This gives analysts a new perspective on the data stored on the blockchain.
• Authors: Jean-Daniel Fekete, Petra Isenberg, Christoph Kinkeldey
• Contacts: Jean-Daniel Fekete, Petra Isenberg

7.1.3 PAOHvis

• Name: Parallel Aggregated Ordered Hypergraph Visualization
• Keywords: Dynamic networks, Hypergraphs
• Functional Description: Parallel Aggregated Ordered Hypergraph (PAOH) is a novel technique to visualize dynamic hypergraphs 29. Hypergraphs are a generalization of graphs where edges can connect more than two vertices. Hypergraphs can be used to model co-authorship networks with multiple authors per article, or networks of business partners. A dynamic hypergraph evolves over discrete time slots. A PAOH display represents vertices as parallel horizontal bars and hyperedges as vertical lines that connect two or more vertices. We believe that PAOH is the first technique with a highly readable representation of dynamic hypergraphs without overlaps. It is easy to learn and is well suited for medium size dynamic hypergraph networks such as those commonly generated by digital humanities projects - our driving application domain (see Fig. 2).
• URL: https://aviz.fr/paohvis
• Contacts: Paola Tatiana Llerena Valdivia, Jean-Daniel Fekete

7.1.4 AR Collaborative Visualization

• Name: AR Collaborative Visualization
• Keywords: Augmented reality, Collaborative science, Android
• Functional Description: Allows to look at VTK datasets using AR-HMD (Microsoft HoloLens) in multi-users environments (i.e., one headset per user). A Multi-touch tablet is provided per user to manipulate the environment.
• Contact: Tobias Isenberg

10.1.1 Inria associate team not involved in an IIL

10.1.2 Inria international partners

10.2.1 Collaborations in European programs, except FP7 and H2020

• Program: ANR PRCI
• Project acronym: MicroVis
• Project title: Micro visualizations for pervasive and mobile data exploration
• Duration: 11/2019 – 08/2022
• Coordinator: Petra Isenberg
• Other partners: University of Stuttgart
• Abstract: The goal of this joint Franco-German project is to study very small data visualizations, micro visualizations, in display contexts that can only dedicate minimal rendering space for data representations. We will study human perception of and interaction with micro visualizations given small as well as complex data. The increasing demand for data visualizations on small mobile devices such as fitness tracking armbands, smart watches, or mobile phones drives our research. Given this usage context, we focus on situations in which visualizations are used “on the go,” while walking, riding a vehicle, or running. It is still unclear to which extent our knowledge of desktop-sized visualizations transfers to contexts that involve minimal display space, diverse viewing angles, and moving displays.

• Program: 2016 FWF–ANR Call for French-Austrian Joint Projects
• Project acronym: ILLUSTRARE
• Project title: Integrative Visual Abstraction of Molecular Data
• Duration: 48 months, extended until the end of June 2021
• Coordinator: Tobias Isenberg and Ivan Viola
• Other partners: TU Wien, Austria; King Abdullah University of Science and Technology, Saudi-Arabia
• Abstract: The essential building block of visualization is the phenomenon of visual abstraction. While visual abstraction is intuitively understood, there is no scientific theory associated with it that would be useful in the visualization synthesis process. Our central aim of this project is thus to gain better understanding of the visual abstraction characteristics. We lay down a hypothetical initial basis of theoretical foundations of visual abstractions in the proposal. We hypothesize that visual abstraction is a multidimensional phenomenon that can be spanned by axes of abstraction. Besides abstractions associated with a static structure we take a closer look at abstractions related to dynamics, procedures, and emergence of the structure. We also study abstraction characteristics related to multi-scale phenomena defined both in space and in time. This hypothetical basis is either supported or rejected by means of exemplary evidence from the specific application domain of structural biology. Structural biology data is very complex, it includes the aspect of emergence and it is defined over multiple scales. Furthermore, abstraction has led to key discoveries in biology, such as the organization of the DNA. We study the multiscale visual abstraction characteristics on the visualization of long nucleic strands and the abstractions that convey emerging phenomena on visualization of molecular machinery use cases. From these two fields we work toward a theory of visual abstraction in a bottom-up manner, investigating the validity of the theory in other application domains as well.

• Program: CHIST-ERA
• Project acronym: IVAN
• Project title: Interactive and Visual Analysis of Networks
• Duration: May 2018 - April 2021
• Coordinator: Dr. Torsten Möller, Uni Wien, Austria
• Other partners: EPFL, Switzerland, Inria France, Uni Wien, Austria
• Abstract:

  The main goal of IVAN is to create a visual analysis system for the exploration of dynamic or time-dependent networks (from small to large scale). Our contributions will be in three principal areas:

  novel algorithms for network clustering that are based on graph harmonic analysis and level-of-detail methods;

  the development of novel similarity measures for networks and network clusters for the purpose of comparing multiple network clusterings and the grouping (clustering) of different network clusterings; and

  a system for user-driven analysis of network clusterings supported by novel visual encodings and interaction techniques suitable for exploring dynamic networks and their clusterings in the presence of uncertainties due to noise and uncontrolled variations of network properties.

  Our aim is to make these novel algorithms accessible to a broad range of users and researchers to enable reliable and informed decisions based on the network analysis.

Member of the organizing committees

• Pierre Dragicevic: alt.chi co-chair at ACM CHI 2021.
• Petra Isenberg: Member of the IEEE Visualization reVISe Committee

Member of Conference Program Committees

• Jean-Daniel Fekete: Eurovis 2020
• Petra Isenberg: IEEE InfoVis 2020
• Tobias Isenberg: IEEE SciVis 2020, BELIV 2020

Reviewer

• Tobias Isenberg: ISMAR 2020, IEEE SciVis 2020, UIST 2020
• Petra Isenberg: ACM CHI 2020, EuroVA 2020, EuroVis 2020, MobileHCI 2020

Too many to mention.

11.1.1 Journal

• Jean-Daniel Fekete: Member of the Publication Board of Eurographics
• Jean-Daniel Fekete: Member of the Scientific Committee of the French journal "Humanités Numériques"

11.1.2 Invited talks

• Jean-Daniel Fekete: Keynote of the Chinese visual computing summer school, "Progressive data analysis: a new language paradigm for scalability in exploratory data analysis", virtual, July 17, 2020
• Jean-Daniel Fekete: Seminar at Inria Saclay, "Demandez le programme", "Progressivis", Palaiseau, Feb. 27, 2020
• Petra Isenberg: Vis Evaluation – Moving into the Next Decade; Panel statement given at the Beliv Workshop 2020, held virtually; October 2020
• Petra Isenberg: When Visualization meets HCI; Keynote at the Mensch und Computer 2020 Conference (held virtually); September 2020

11.1.3 Leadership within the scientific community

• Tobias Isenberg: Steering Committee for ACM/EG Expressive; Steering Committee for BELIV workshop series
• Petra Isenberg: Vice-chair of the IEEE Visualization Steering Committee; Steering Committee Member for BELIV workshop series

11.1.4 Research administration

• Pierre Dragicevic: member of the CCSU (Commission Consultative de Spécialistes de l'Université Paris-Saclay)
• Pierre Dragicevic: member of the Conseil de Labo LISN
• Pierre Dragicevic: co-chair of the IID axis of the Labex Digicosme
• Pierre Dragicevic: member of the CER (Comité d'Éthique de la Recherche) Paris-Saclay
• Petra Isenberg: Member of the Commission de Developpement Technologique, Inria

11.2.1 Teaching

• Master: Petra Isenberg, Interactive Information Visualization, 21h en equivalent, niveau (M1, M2), Univeristé Paris-Saclay, France.
• Master: Petra Isenberg, Graphisme et Visualisation, 24h en equivalént TD, niveau (M2), Polytech Paris-Saclay, France.
• Master: Petra Isenberg, Visual Analytics, 48h equivalent, niveau (M2), CentraleSupelec, France.
• Licence : Tobias Isenberg, “Introduction to Computer Graphics”, 18h en équivalent TD, L3, Polytech Paris-Saclay, France.
• Master: Tobias Isenberg, “Photorealistic Rendering/Advanced Computer Graphics”, 28h en équivalent TD, M2, Polytech Paris-Saclay, France.
• Master: Tobias Isenberg, “Illustrative Visualization”, 6h en équivalent TD, M2, ENSTA, France.
• Master: Jean-Daniel Fekete, Scalability in Visual Analytics, 3h, niveau (M2), CentraleSupelec, France.
• Training: Jean-Daniel Fekete: Crash course on HCI and Visualization, IRT SystemX, 6h

11.2.2 Supervision

• PhD in progress: Marie Destandau, Path-Based Interactive Visual Exploration of Knowledge Graphs, Université Paris-Saclay, defense planned for December 2020, co-supervised by Emmanuel Pietriga and Jean-Daniel Fekete.
• PhD in progress: Alexis Pister, Exploration, analyse, interprétabilité de larges réseaux historiques, Université Paris-Saclay, defense planned for 2022, co-supervised by Jean-Daniel Fekete and Christophe Prieur, Telecom Paris.
• PhD in progress: Jiayi Hong, Interactive Visualization and Classification of Temporal Developments in 3D Datasets, Univ. Paris-Saclay, defense planned for December 2022, Tobias Isenberg.
• PhD in progress: Mickaël Sereno, Collaborative Data Exploration and Discussion Supported by AR, Univ. Paris-Saclay, defense planned for December 2021, Tobias Isenberg.
• PhD completed: Xiyao Wang, Augmented Reality Environments for the Interactive Exploration of 3D Data, Univ. Paris-Saclay, defended successfully December 16, 2020, Tobias Isenberg.
• PhD completed: Sarkis Halladjian, Spatially Integrated Abstraction of Genetic Molecules, Univ. Paris-Saclay; defended successfully December 14, 2020, Tobias Isenberg.
• PhD in progress: Mohamad Alaul Islam: MicroVisualization for mobile and ubiquitous data exploration, Université Paris-Saclay, defense planned for 2022, Petra Isenberg and Jean-Daniel Fekete.
• PhD in progress: Natkamon Tovanich: Blockchain Visual Analytics, defense planned for 2021, Université Paris-Saclay, Petra Isenberg and Jean-Daniel Fekete.
• PhD in progress: Lijie Yao: Visualization in Motion, defense planned for 2023, Université Paris-Saclay, Petra Isenberg and Anastasia Bezerianos.

11.2.3 Juries

• Jean-Daniel Fekete: Vincent Raveneau, "Interaction in Progressive Visual Analytics. An application to progressive sequential pattern mining", Université de Nantes, Nov. 4, 2020
• Jean-Daniel Fekete: Anastasia Bezerianos, HdR, "Increasing the bandwidth of interactive visualizations using complex display environments and targeted designs", Université Paris-Saclay, Jul. 2nd, 2020
• Jean-Daniel Fekete: Robin Cura, "Modéliser des systèmes de peuplement en interdisciplinarité. Co-construction et exploration visuelle d'un modèle de simulation", Université Paris 1, Mar. 6, 2020
• Petra Isenberg: Paul Lapides, “Information Visualization for Exploration and Self-Reflection in Social Media”, University of Calgary, 2020
• Petra Isenberg: Fern Bishop, “Investigating the Support of Visualization Creation Activities with Children” University of St. Andrews, 2020

11.3.1 Articles and contents

• Guest post by Steve Haroz on “Statistical Modeling, Causal Inference, and Social Science" — "IEEE’s Refusal to Issue Corrections”. https://statmodeling.stat.columbia.edu/2020/12/10/ieees-refusal-to-issue-corrections/
• Guest post by Steve Haroz on “Statistical Modeling, Causal Inference, and Social Science" — "Update on IEEE’s refusal to issue corrections”. https://statmodeling.stat.columbia.edu/2020/12/23/update-on-ieees-refusal-to-issue-corrections/
• Article “Arkangel dans Black Mirror : la réalité diminuée” by Lonni Besançon, Amir Semmo, Tobias Isenberg, and Pierre Dragicevic in Interstices 57https://interstices.info/arkangel-dans-black-mirror-la-realite-diminuee/
• Blog post on “Le divulgâcheur Arkangel de Black Mirror” in Le Monde's binaire blog about work by Lonni Besançon, Amir Semmo, Tobias Isenberg, and Pierre Dragicevic 10https://www.lemonde.fr/blog/binaire/2020/04/14/le-divulgacheur-arkangel-de-black-mirror/
• Blog post on “Can Graphic Content Inform Without the Shock-factor?” in Wiley's researcher blog series about work by Lonni Besançon, Amir Semmo, Tobias Isenberg, and Pierre Dragicevic 10https://www.wiley.com/network/researchers/researcher-blogs/can-graphic-content-inform-without-the-shock-factor
• Pierre Dragicevic and Yvonne Jansen: the data physicalization wiki (http://dataphys.org/) and the List of Physical Visualizations and Related Artefacts (http://dataphys.org/list/, 600 weekly visits) are continuously being updated.

11.3.2 Education

• Petra Isenberg: Interview for the Linkedin Data visualization Series

International journals

• 10 articleLonniL. Besançon, AmirA. Semmo, David J.D. Biau, BrunoB. Frachet, VirginieV. Pineau, El HadiE. Sariali, MarcM. Soubeyrand, RabahR. Taouachi, TobiasT. Isenberg and PierreP. Dragicevic. Reducing Affective Responses to Surgical Images and Videos Through StylizationComputer Graphics Forum391January 2020, 462--483
  HALDOIback to textback to textback to text
• 11 articleLonniL. Besançon, AndersA. Ynnerman, Daniel FD. Keefe, LingyunL. Yu and TobiasT. Isenberg. The State of the Art of Spatial Interfaces for 3D VisualizationComputer Graphics Forum401February 2021, 293--326
  HALDOI
• 12 articleMatthewM. Brehmer, BongshinB. Lee, PetraP. Isenberg and Eun KyoungE. Choe. A Comparative Evaluation of Animation and Small Multiples for Trend Visualization on Mobile PhonesIEEE Transactions on Visualization and Computer Graphics2612020, 364--374
  HALDOI
• 13 article PhilippeP. Caillou, JonasJ. Renault, Jean-DanielJ.-D. Fekete, Anne-CatherineA.-C. Letournel and MichèleM. Sebag. Cartolabe: A Web-Based Scalable Visualization of Large Document Collections IEEE Computer Graphics and Applications 2020
  HAL DOI
• 14 article JianJ. Chen, MengM. Ling, RuiR. Li, PetraP. Isenberg, TobiasT. Isenberg, MichaelM. Sedlmair, TorstenT. Möller, Robert SR. Laramee, Han-WeiH.-W. Shen, KatharinaK. Wünsche and QiruQ. Wang. VIS30K: A Collection of Figures and Tables from IEEE Visualization Conference Publications IEEE Transactions on Visualization and Computer Graphics 2021
  HAL DOI
• 15 articleAndyA. Cockburn, PierreP. Dragicevic, LonniL. Besançon and CarlC. Gutwin. Threats of a replication crisis in empirical computer scienceCommunications of the ACM638August 2020, 70-79
  HALDOIback to text
• 16 articleMarieM. Destandau and Jean-DanielJ.-D. Fekete. The Missing Path: Diagnosing Incompleteness in Linked DataInformation Visualization201February 2021, 66-82
  HALDOI
• 17 articleEvanthiaE. Dimara, StevenS. Franconeri, CatherineC. Plaisant, AnastasiaA. Bezerianos and PierreP. Dragicevic. A Task-based Taxonomy of Cognitive Biases for Information VisualizationIEEE Transactions on Visualization and Computer Graphics2622020, 1413 - 1432
  HALDOI
• 18 articleJean-DanielJ.-D. Fekete and JulianaJ. Freire. Exploring Reproducibility in VisualizationIEEE Computer Graphics and Applications405September 2020, 108-119
  HALDOI
• 19 articleJohannesJ. Fuchs, PetraP. Isenberg, AnastasiaA. Bezerianos, MatthiasM. Miller and DanielD. Keim. Teaching Clustering Algorithms With EduClust: Experience Report and Future DirectionsIEEE Computer Graphics and Applications402March 2020, 98--102
  HALDOI
• 20 articleSarkisS. Halladjian, HaichaoH. Miao, DavidD. Kouřil, EduardE. Gröller, IvanI. Viola and TobiasT. Isenberg. ScaleTrotter: Illustrative Visual Travels Across Negative ScalesIEEE Transactions on Visualization and Computer Graphics261January 2020, 654-664
  HALDOI
• 21 articleSteveS. Haroz. The Set Size Bias in Ensemble Comparison (Or Why Showing Raw Data May Be Misleading)Journal of Vision2011October 2020, 741
  HALDOI
• 22 articleJaeminJ. Jo, JinwookJ. Seo and Jean-DanielJ.-D. Fekete. PANENE: A Progressive Algorithm for Indexing and Querying Approximate k-Nearest NeighborsIEEE Transactions on Visualization and Computer Graphics262February 2020, 1347-1360
  HALDOIback to textback to textback to textback to textback to text
• 23 article DavidD. Kouřil, TobiasT. Isenberg, BarboraB. Kozlíková, MiriahM. Meyer, EduardE. Gröller and IvanI. Viola. HyperLabels---Browsing of Dense and Hierarchical Molecular 3D Models IEEE Transactions on Visualization and Computer Graphics 2021
  HAL DOI
• 24 articleBongshinB. Lee, Eun KyoungE. Choe, PetraP. Isenberg, KimK. Marriott and John T.J. Stasko. Reaching Broader Audiences With Data VisualizationIEEE Computer Graphics and Applications4022020, 82--90
  HAL
• 25 article YuchengY. Lu, LuyuL. Cheng, TobiasT. Isenberg, Chi-WingC.-W. Fu, GuoningG. Chen, HuiH. Liu, OliverO. Deussen and YunhaiY. Wang. Curve Complexity Heuristic KD-trees for Neighborhood-based Exploration of 3D Curves Computer Graphics Forum 40 2 May 2021
  HAL
• 26 article LuizL. Morais, YvonneY. Jansen, NazarenoN. Andrade and PierreP. Dragicevic. Showing Data about People: A Design Space of Anthropographics IEEE Transactions on Visualization and Computer Graphics 2020
  HAL back to text back to text
• 27 articleAlexisA. Pister, PaoloP. Buono, Jean-DanielJ.-D. Fekete, CatherineC. Plaisant and PaolaP. Valdivia. Integrating Prior Knowledge in Mixed Initiative Social Network ClusteringIEEE Transactions on Visualization and Computer Graphics272February 2021, 1775 - 1785
  HALDOIback to text
• 28 article MickaelM. Sereno, XiyaoX. Wang, LonniL. Besançon, Michael JM. Mcguffin and TobiasT. Isenberg. Collaborative Work in Augmented Reality: A Survey IEEE Transactions on Visualization and Computer Graphics 2021
  HAL DOI
• 29 articlePaola RP. Valdivia, PaoloP. Buono, CatherineC. Plaisant, NicoleN. Dufournaud and Jean-DanielJ.-D. Fekete. Analyzing Dynamic Hypergraphs with Parallel Aggregated Ordered Hypergraph VisualizationIEEE Transactions on Visualization and Computer Graphics271January 2021, 1-13
  HALDOIback to textback to textback to text
• 30 article AoyuA. Wu, WaiW. Tong, TimT. Dwyer, PetraP. Isenberg, BongshinB. Lee and HuaminH. Qu. MobileVisFixer: Tailoring Web Visualizations for Mobile Phones Leveraging an Explainable Reinforcement Learning Framework IEEE Transactions on Visualization and Computer Graphics 2021
  HAL DOI

International peer-reviewed conferences

• 31 inproceedingsLeilaniL. Battle, PhilippP. Eichmann, MarcoM. Angelini, TizianaT. Catarci, GiuseppeG. Santucci, YukunY. Zheng, CarstenC. Binnig, Jean-DanielJ.-D. Fekete and DominikD. Moritz. Database Benchmarking for Supporting Real-Time Interactive Querying of Large DataSIGMOD ’20 - International Conference on Management of DataProceedings of the 2020 ACM SIGMOD International Conference on Management of DataPortland, OR, United Stateshttps://sigmod2020.org/June 2020, 1571-1587
  HALDOI
• 32 inproceedingsTanjaT. Blascheck and PetraP. Isenberg. A Replication Study on Glanceable Visualizations: Comparing Different Stimulus Sizes on a Laptop ComputerProceedings of the Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3 IVAPPVienna, Austria2021, 133-143
  HALDOI
• 33 inproceedings ReneR. Cutura, MichaëlM. Aupetit, Jean-DanielJ.-D. Fekete and MichaelM. Sedlmair. Comparing and Exploring High-Dimensional Data with Dimensionality Reduction Algorithms and Matrix Visualizations AVI' 20 - International Conference on Advanced Visual Interfaces Ischia Island, Italy September 2020
  HAL DOI
• 34 inproceedings PascalP. Goffin, PetraP. Isenberg, TanjaT. Blascheck and WesleyW. Willett. Interaction Techniques for Visual Exploration Using Embedded Word-Scale Visualizations CHI '20: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems CHI '20 - 38th SIGCHI conference on Human Factors in computing systems Proceedings of the Conference on Human Factors in Computing Systems (CHI) Honolulu, United States April 2020
  HAL DOI
• 35 inproceedings JiayiJ. Hong, FerranF. Argelaguet Sanz, AlainA. Trubuil and TobiasT. Isenberg. Design and Evaluation of Three Selection Techniques for Tightly Packed 3D Objects in Cell Lineage Specification in Botany Graphics Interface Mississauga, Canada May 2021
  HAL
• 36 inproceedings AlaulA. Islam, AnastasiaA. Bezerianos, BongshinB. Lee, TanjaT. Blascheck and PetraP. Isenberg. Visualizing Information on Watch Faces: A Survey with Smartwatch Users Short Papers of the IEEE Conference on Visualization (VIS) IEEE VIS 2020 - Visualization Conference Short Papers Short Papers of the IEEE Conference on Visualization (VIS) Los Alamitos, CA, United States October 2020
  HAL DOI
• 37 inproceedings LuizL. Morais, YvonneY. Jansen, NazarenoN. Andrade and PierreP. Dragicevic. Can Anthropographics Promote Prosociality? A Review and Large-Sample Study CHI 2021 - Conference on Human Factors in Computing Systems Yokohama / Virtual, Japan https://chi2021.acm.org/ May 2021
  HAL DOI back to text back to text
• 38 inproceedings Best paper ChatchavanC. Wacharamanotham, LukasL. Eisenring, SteveS. Haroz and FlorianF. Echtler. Transparency of CHI Research Artifacts: Results of a Self-Reported Survey CHI Conference on Human Factors in Computing Systems Event not held due to COVID, Unknown Region 2020
  HAL DOI back to text
• 39 inproceedings Best paper LiwenhanL. Xie, JamesJ. O'Donnell, BenjaminB. Bach and Jean-DanielJ.-D. Fekete. Interactive Time-Series of Measures for Exploring Dynamic Networks AVI' 20 - International Conference on Advanced Visual Interfaces 20 Ischia Island, Italy September 2020
  HAL DOI back to text

Conferences without proceedings

• 40 inproceedings GennadyG. Andrienko, NataliaN. Andrienko, Steven M.S. Drucker, Jean-DanielJ.-D. Fekete, DanyelD. Fisher, StavrosS. Idreos, TimT. Kraska, GuoliangG. Li, Kwan-LiuK.-L. Ma, Jock DJ. Mackinlay, AnttiA. Oulasvirta, TobiasT. Schreck, HeidrunH. Schmann, MichaelM. Stonebraker, DavidD. Auber, NikosN. Bikakis, Panos KP. Chrysanthis, GeorgeG. Papastefanatos and Mohamed AM. Sharaf. Big Data Visualization and Analytics: Future Research Challenges and Emerging Applications BigVis 2020 - 3rd International Workshop on Big Data Visual Exploration and Analytics Workshop Proceedings of the EDBT/ICDT 2020 Joint Conference Copenhagen, Denmark http://ceur-ws.org/Vol-2578/ March 2020
  HAL
• 41 inproceedings MarieM. Destandau and Jean-DanielJ.-D. Fekete. Diagnosing Incompleteness in Wikidata with The Missing Path Wiki Workshop 2020 Taipei, Taiwan April 2020
  HAL
• 42 inproceedings NatkamonN. Tovanich, NicolasN. Soulié and PetraP. Isenberg. Visual analytics of bitcoin mining pool evolution : on the road toward stability? 3rd International Workshop on Blockchains and Smart Contracts (BSC 2020-2021), held in conjunction with the 11th IFIP International Conference on New Technologies, Mobility and Security (IFIP NTMS 2021) Paris, France 2021
  HAL
• 43 inproceedings XiyaoX. Wang, DavidD. Rousseau, LonniL. Besançon, MickaelM. Sereno, MehdiM. Ammi and TobiasT. Isenberg. Towards an Understanding of Augmented Reality Extensions for Existing 3D Data Analysis Tools CHI '20: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems CHI '20 - 38th SIGCHI conference on Human Factors in computing systems Honolulu, United States April 2020
  HAL DOI

Scientific book chapters

• 44 inbook PierreP. Dragicevic, YvonneY. Jansen and AndrewA. Vande Moere. Data Physicalization Springer Handbook of Human Computer Interaction Springer Reference 2021
  HAL back to text
• 45 inbookIvanI. Viola, MinM. Chen and TobiasT. Isenberg. Visual AbstractionFoundations of Data VisualizationAugust 2020, 15--37
  HALDOI

Reports & preprints

• 46 misc MarieM. Destandau, OlivierO. Corby, Jean-DanielJ.-D. Fekete and AlainA. Giboin. Path Outlines: Browsing Path-Based Summaries of Linked Datasets February 2020
  HAL
• 47 reportPierreP. Dragicevic. A Mean Difference is an Effect SizeInria Saclay Ile de FranceJuly 2020, 1-12
  HAL
• 48 misc ChristophC. Kinkeldey, Jean-DanielJ.-D. Fekete, TanjaT. Blascheck and PetraP. Isenberg. Visualizing and Analyzing Entity Activity on the Bitcoin Network March 2020
  HAL
• 49 misc DavidD. Kouřil, OndřejO. Strnad, PeterP. Mindek, SarkisS. Halladjian, TobiasT. Isenberg, Eduard M.E. Gröller and IvanI. Viola. Molecumentary: Scalable Narrated Documentaries Using Molecular Visualization August 2020
  HAL

Other scientific publications

• 50 misc JianJ. Chen, MengM. Ling, RuiR. Li, PetraP. Isenberg, TobiasT. Isenberg, MichaelM. Sedlmair, TorstenT. Möller, RobertR. Laramee, Han-WeiH.-W. Shen, KatharinaK. Wünsche and QiruQ. Wang. IEEE VIS Figures and Tables Image Dataset July 2020
  HAL DOI
• 51 misc StéphaneS. Gosset, MickaelM. Sereno, LonniL. Besançon and TobiasT. Isenberg. Tangible Volumetric Brushing in Augmented Reality Salt Lake City, United States October 2020
  HAL
• 52 misc JiayiJ. Hong, FerranF. Argelaguet, AlainA. Trubuil and TobiasT. Isenberg. An Interactive System for Analyzing Plant Embryo Cell Division Salt Lake City, United States October 2020
  HAL
• 53 misc MickaelM. Sereno and TobiasT. Isenberg. Subjective Views in Co-Located Augmented Reality-Initial Design Salt Lake City, United States October 2020
  HAL
• 54 misc NatkamonN. Tovanich, AlexisA. Pister, GaëlleG. Richer, PaolaP. Valdivia, Jean-DanielJ.-D. Fekete, ChristopheC. Prieur and PetraP. Isenberg. GraphletMatchMaker: Visual Analytics Approaches to Graph Matching in Cybersecurity Communities Virtual Conference, United States October 2020
  HAL
• 55 misc LijieL. Yao, AnastasiaA. Bezerianos and PetraP. Isenberg. Situated Visualization in Motion Salt Lake City, United States October 2020
  HAL
• 56 misc YuanyangY. Zhong, TobiasT. Isenberg and PetraP. Isenberg. Black-and-White Textures for Visualization on E-ink Displays Salt Lake City, United States October 2020
  HAL

Scientific popularization

• 57 article L. Besançon, A. Semmo, T. Isenberg and P. Dragicevic. Arkangel dans Black Mirror : la réalité diminuée Interstices April 2020
  HAL back to text