Visual analytics: A human in the Loop

Talks & interactive workshop, September 30, BC 420

When people make decisions using the results of the machine learning algorithms that power AI systems, it’s vitally important that they are able to interpret and understand the results correctly. These are often communicated in the form of data visualizations and can sometimes be very complex to give meaning to. The central research question in visual analytics is how to design data visualisations that can be correctly understood by users to help them to make the right decisions. This question lies in the intersection of human-computer interaction, machine learning and data visualization.

This short workshop will include 4 talks and individual meetings with the speakers.


09:15 : Dr. Wafa Johal, University of Melbourne

10:00 : Dr. Niklas Elmqvist, University of Maryland 

10:45 : Coffee break

11:15 : Dr. Evanthia Dimara, Utrecht University

12:00 : Dr. Benjamin Bach, University of Edinburgh

12:45 : End


Hall BC 420, EPFL


While early research in Human Robot Interaction (HRI) focused on teleoperation and human supervision of robots performing tasks, the developments in sensing, actuation and decision making, have enabled robots to cary some tasks autonomously. As a consequence of having more autonomous robots operating collaboratively with humans, data visualisation has been an emerging topic in HRI. 

This presentation will present a short literature background of data visualisation for HRI. We will look at a couple of concrete examples in robots for learning and in explainable robotics. Finally we will discuss opportunities for merging visual analytics and Human-Robot Interaction to allow human users to better understand autonomous robots.

Mobile computing, virtual and augmented reality, and the internet of things (IoT) have transformed the way we interact with computers. Artificial intelligence and machine learning have unprecedented potential for amplifying human abilities. But how have these technologies impacted data analysis, and how will they cause data analysis to change in the future?

In this talk, I will review my group’s sustained efforts of going beyond the mouse and the keyboard into the “metaverse” of analytics: large-scale, distributed, ubiquitous, immersive, and increasingly mobile forms of data analytics augmented and amplified by AI/ML models. I will also present my vision for the fundamental theories, applications, design studies, technologies, and frameworks we will need to fulfill the vast potential of this exciting new area in the future.

How can organizations cultivate a culture in which decisions are based on complex data and scientific evidence, if the current most accessible tools that decision makers use are a paper, a pencil and a spreadsheet? Unaided by technology, human intuition is overloaded with data or hindered by systematic biases. Fully automated technology, such as state‐of‐the‐art Artificial Intelligence (AI) solutions, is not yet suitable for ill‐defined problems.

Somewhere in between is Information Visualization (Infovis), a field that studies the design of effective data interfaces which keep the user in control of the reasoning process while helping them deal with ill‐defined problems. Yet, despite the rich variety of visual analytics software, real‐world decision makers have not fully exploited the benefits of visualizations. Possible explanations for their lack of adoption can be that decision makers are not data literate or simply not aware of more powerful existing solutions. An alternative explanation could be that visualizations are simply not well-designed to facilitate the creative act of decision making.

This presentation leverages the second explanation by identifying challenges and opportunities for novel visualization tools.  Our findings from an empirical survey and interviews with people who make organizational decisions stress the need to expand visualization design beyond data analysis objectives into tools for information management. The related publications are openly available here

This talk reflects on supporting people in using, creating, and designing (interactive) data visualizations and visual analytics systems. With the increasing prominence of data visualization in public and professional life comes the need for education in data and visualization literacy to prevent misinformation and empower people in using and creating (interactive) visualizations.

Based on a set of case studies from my previous research—interactive visualization for the analysis of multivariate and temporal networks (e.g.,; visualization for temporal data; dashboard design; visualization and interaction in immersive environments; and data comics for storytelling and education—this talk motivates and demonstrates recent work by the VisHub Lab on supporting people without specific visualization training in efficiently using and creating visualisations. We’ll discuss barriers in approaching interactive visualisations as well as solutions such as visual cheatsheets, design patterns, tool collections, audience participation, or workshops. The talk invites for a discussion on challenges in visualization design and usage. 


Dr. Wafa Johal

Data Visualisation for Human-Robot Interaction

Dr. Wafa Johal is a Senior Lecturer in Computer Sciences at the Faculty of Engineering and Information Technology at the University of Melbourne. Previously, she was a Lecturer at the University of New South Wales, Sydney. From 2015 to 2019, she was a researcher at the Computer-Human Interaction Lab for Learning and Instruction and the Mobots Group at EPFL, Switzerland.

Professor Niklas Elmqvist

Anytime Anywhere All At Once: Data Analytics in the Metaverse

Niklas Elmqvist is a full professor in the iSchool (College of Information Studies) at University of Maryland, College Park. He received his Ph.D. in computer science in 2006 from Chalmers University in Gothenburg, Sweden.

Assistant Professor Evanthia Dimara

Opportunities and Challenges for Data Visualization to Support Decision Making

Evanthia Dimara is an Assistant Professor of Information Visualization at Utrecht University in the Visualization and Graphics group (VIG). Her fields of research are Information Visualisation and Human-Computer Interaction. 

Associate Professor Benjamin Bach

Visualization Empowerment: Towards Tools and Methods for Creating Effective Data Visualizations

Benjamin is a Reader (Associate Professor) in Design Informatics and Visualization at the School of Informatics, University of Edinburgh. He founded and is now co-leading the VisHub Lab as part of the Institute for Design Informatics.