Visualization Empowerment:
How to Teach and Learn Data Visualization

After introducing the seminar and presenting various organizational matters, the first day started with a sketching activity organized by Tatiana Losev from Simon Fraser University in Vancouver, Canada (Section 6.1). This activity acted as an icebreaker for participants start to get to know each other, and to initiate reflections on the seminar topic in a playful way. Participants were invited to sketch their relationship to the seminar topic, and to then use this sketch to introduce themselves in the form of a 1-minute presentation. Some participants sketched their relation to teaching and research in a literal, metaphorical or abstract way, others focused on the variety of themes and questions to be discussed, and again others created visual representations of their past curricula (see Figure 2). The activity brought to the fore an exciting diversity of viewpoints as well as the coverage of interest toward the topics of the seminar.

The sketching activity was followed by a brainstorming session that invited participants to identify high- and low-level topics around visualization empowerment and teaching, that they wanted to discuss during the seminar. This session was intended to initiate and fuel discussions that would take place in the form of smaller working groups throughout the week. Participants noted topics on sticky notes that we then collaboratively reviewed and grouped (see Figure 3). We identified a great diversity of themes including design creativity, physicalization, ethics, democratization, scalability of teaching, humanism, tools, hybrid and online teaching, teaching methods, community building, success stories and inspiration, learning goals, planning, contexts, and barriers, audiences (from practitioners to children), evaluation and assessment, measuring learning progress, interdisciplinarity, critical thinking, inclusivity, resources, and cataloging educational material.

We ranked topics in a voting activity, based on participants’ interests to discuss them. This led to the formation of initially four working groups: teaching methods, democratization, creativity, and physicalization.

• $\mathrm{■}$

  The Teaching Methods Group (see Section 4.5) first focused on the diversity of challenges in visualization education. They explored possibilities of formalizing a multidimensional problem space to capture these challenges.

• $\mathrm{■}$

  The Democratization Group (see Section 4.3) discussed practices, believes, intentions and biases that influence visualization teaching and creating with the aim to make visualization more accessible. Based on these discussions they decided to question the value behind our teaching activities.

• $\mathrm{■}$

  The Creativity Group (see Section 4.1) focused on how to teach creativity and criticality in visualization; how one can be creative in teaching visualization, how one can teach creativity thought visualization, and eventually focused on an activity book for novice visualization designers.

• $\mathrm{■}$

  The Physicalization Group (see Section 4.4) discussed how data physicalizations could be a mediator for teaching and learning activities, but also how it can be used to breach disciplines, and also how inclusive the physicalization can be for teaching and learning.

These working groups continued discussions throughout the week in different participant constellations. A number of participants shifted between groups to absorb different discussions, which proved to be useful for cross-dissemination across working groups.

At the end of the day, i.e., after some initial discussions and topic finding within the individual groups, each group briefed the entire seminar on their discussion and focus. The North America group joined to get updated on the European groups.

This day was mainly reserved for discussions within the individual working groups. “Over night”, the one organizer based in North America lead the North America discussion group which decided on the topic of Improvisation in visualization and what can be learned from improvisation in art for how to approach visualization (design). We started the day with a short briefing into the day schedule and asked people if they wanted to switch or split groups. Then, we had a series of short 5min talks from seminar participants Andrew Manches (education) providing a learning science perspective in his talk The potential of a more embodied approach to supporting children’s data understanding (Section 5.1); Andy Kirk (freelance visualization designer and educator), provided a non academic visualization trainer perspective “How I can help you? How can you help me?”(Section 5.2); eventually Peter Cheng brought a cognitive psychology perspective through his talk “Cognitive Science of Representational Systems” (Section 5.3). A joined question and answer session followed these talks. Then, participants broke out into their groups. At the beginning of the afternoon Wesley Willet ran a visualization activity “Visualization future card game”(Section 6.2).

Again, at the end of the day, all working groups met, including the North America group on ImproVISation to brief the other seminar participants.

We started the day, again, with four short 5min talks: Isabel Meirelles (visualization design) “Breaking the Monolith” (Section 5.4); Jason Dykes (cartography and visualization) about assessment of visualization teaching (Section 5.5); Jonathan Roberts (visualization) “From visioning to solution via sketching” (Section 5.6). Following these presentations a vibrant and spontaneous conversation happened among participants about the different assessment strategies in various teaching constraints. The discussion was so spontaneous and interesting it took most of the morning.

Before the afternoon socialization, Samuel Huron ran a visualization activity with all participants, aiming to classify visualization activities. Prior to the seminar, Samuel had invited participants to submit short descriptions of activities they do in their classes with their students. This collection was printed on small cards, one activity per card, and distributed among each working group. Each working group came up with a different classification scheme which is currently informing an ongoing discussion. There was no evening briefing with the North America group due to the socialization activity.

During the morning talks, Fateme Rajabiyazdi shared some of the lessons learned in her class during the talk “Teaching visualization Free Form ”(Section 5.8). Doris Kosminsky discuss how can we empower mother with health data in Brazil in her talk “Reflections on learning and empowerment of those represented in health visualization”. Then Alexandra Diehl presented how Visguides ⁵⁵5https://visguides.org/ could be use for education. Visguides is an open community project to provide guidance and support on visualization design in a web forum (Section 5.7). Last Till Nagel presented the goal, reflect on the process and the design of the book Making with data (Section 5.9).

Mandy Keck proposed to create a pre-approved symposium at IEEE VIS conference, the major international forum for visualization with over 1000 attendees. The symposium would become a major outlet and forum for research around visualization education and learning. It would be a place for research, reflection, creation, and discussion of learning / teaching but also discussing higher-level issues in regards to human-centered approaches to visualization education and design, and build a permanent forum and community around these topics.

Later in the afternoon, Peter Cheng ran an activity to model the cognitive process of reading a visualization through annotating a visualization and then modeling the different cognitive steps of our reading procedure (Section 6.4).

The remainder of this day (morning, afternoon) was reserved for discussions within the working groups. The organizers encouraged goal-oriented thinking and to list and plan the different outcomes of each working group to be reported in the pre-dinner briefing session. Since some participant had to leave early on Friday morning, we started a general discussion about the individual outcomes of this seminar and how to organize working groups beyond the seminar.

We started the day by synthesizing and presenting on one page all the potential outcomes that were planned by the individual working groups. Then, Jason Dykes gave a short talk to introduce data visualization to an audience (Section 5.10). Then the group started a discussion to reflect on the seminar experience and outcomes. This discussion push up to open two other topics thread, one on grants, and one on writing a paper about the main challenges in information visualization teaching and learning. After the coffee break the remaining participants decided to outline collectively a paper on grand challenges in visualization education, effectively forming a sixth working group at the seminar.

This working group focused on creativity in visualization empowerment. Creativity is the use of people’s imagination to engender original ideas, to make and create something, be inventive, design new ideas or make different designs.

Early Education has over two centuries’ experience of designing materials to help children learn abstract concepts – such as colored rods to help children learn numerical relationships. Yet the representational transparency of these materials depends upon existing knowledge of the learner. Representations often integrate a range of conceptual and cultural metaphors that are often known but taken for granted by adulthood.

Simply looking at colored rods will not enable a child to just “get” numbers. Pedagogy, clearly, is key. This captures many things (e.g., narrative, construction activities); my work focuses on interaction – how adults scaffold children’s interaction and learning with materials. In particular, I attend to the multimodality and bidirectionality of scaffolding. It is not just words but other modes such as facial expression, body posture, gaze and gesture that educators employ. Gestures are notably powerful in their capacity to provide schematic, dynamic, visuo-spatial representations coproduced with speech to bridge communication with our environment. And importantly, scaffolding is two-way: children also employ a range of modes to structure and manipulate the support they need. The importance of multimodality is further accentuated when considering emerging theories of what it means to know (and hence what children have “learnt”). Increasing evidence points to the embodied nature of cognition and how learning involves the internalization of body-based experiences, emphasizing the interwoven nature of emotional, social, physical, and cognitive dimensions. When communicating our understanding, we can activate these prior experiences – evident in the emotions and gestures we commonly produce in explanations (makes for a good observational activity at seminars like Dagstuhl).

Increasing attention to the significance of multimodality in how we think and interact has important implications for design as well as pedagogy. This may be greater recognition of existing activities and media (e.g., physicalization) or the potential for more body-based interaction with digital representations (e.g. tangibles, haptics, gesture recognition). More recently, my work has asked how we can tap into emerging theories of cognition and digital tools to help young children (3 years+) understand concepts of data. This is not just a conceptual challenge – children’s worlds are increasingly datafied- from how we measure their “progress” to the (smart) toys we give them. Here there is much potential. Young children understand and often articulate themselves and their interaction with the world (e.g., how old, noisy, tall, active, sleepy, or happy they are) – hence offering a design and learning opportunity through appropriately representing this personal information. Educators already do – colored rods to compare ages, stickers to quantify good behavior or classroom “noisemeters” to maintain sanity. Experts in the field of visualisation/physicalisation have much potential to create a new generation of embodied designs and activities that build upon this foundation.

In this lightning talk I introduced myself to members of the seminar, especially as I’m a non-academic participant so my work may be less than familiar to most. I presented an outline of how I may be able to help my fellow attendees. I opened with an overview of my “boundary spanning” activities, as a freelancer: I publish via my website and podcast series, I teach (academically) and research, I consult and design, I author and present. Above all, perhaps, I train professionals, outside of academia and across a diverse array of client organization types and industries. I described some of the key objectives and approaches I take to the challenge of teaching and learning, and how delighted I was to observe alignment with the seminar’s theme of “Visualization Empowerment”. Given my activities and career experiences, could I be of service to offer any guidance to others?

I then switched over to introduce some matters of interest that I am particularly keen to learn about over the seminar, and maybe get some help from others. Firstly, listing some of the current challenges I experience in the forum of public training: including carving out distinctions in teaching levels (basic >> advanced) for the same and different cohorts, how to teach the concept of elegance and instill journalistic curiosities. Secondly, and finally, issues that specifically affect training in private/client settings, including how to demonstrate (maybe prove?) success of visualisation in terms like ROI, how to encourage organizational readiness for cultural change, ambitions vs. reproducible pragmatism, and the challenges of educating across such a multi-disciplinary skillset.

My approach to the design of information visualisations, and representational systems more generally, combines cognitive science with the analysis of the conceptual structure of the to-be visualized topic. To obtain theoretical and empirical leverage to formulate and test principles of representation design, I create novel diagrammatic systems for conceptually challenging topics and interactive graphical user-interfaces for information intensive decision-making systems. Generalizing over the creation and evaluation of many such systems, I make four claims: (1) STEM topics should be easy to learn; (2) compared to extant conventional visualizations, factor of 2 improvements in problem solving and learning are feasible when representations effectively re-codify knowledge; (3) knowledge re-codification should attempt to capture the conceptual structure of a topic in the graphical structure of the representation; (4) this yields representations that possess semantic transparency and syntactic plasticity.

The talk invited a conversation about the challenges we encounter preparing students to contribute to data visualization practices. There is an unbalance of how skills are taught across disciplines (sciences, arts, humanities, etc). This unbalance affects education at all levels and disciplines. I would like to suggest that we tailor data visualization pedagogical strategies in a situated manner. For that I proposed a scaffold built around thinking processes positioned along the theoretical-practical axis (7). The thinking processes are derived from the literacies needed for data visualization and dependent on the setting and pedagogical goals of the course and needs of our learners (listed alphabetically): analytical, computational, critical, humanistic, numerical, systems, and visual (7). Educators can use the scaffold to identify areas of focus and whether they will approach through practical and/or theoretical means.

I made passionate plea to design assessments so that the teacher gets the information they need to make the best judgment that they can in the time they have. Few of my colleagues do this. Aim to use as little time as you (teacher) can to assess, in ways that are as efficient as possible. Aim to make it fun. Yes, really FUN and INFORMATIVE – how well are you teaching? What can your students do? Assessment is a creative design exercise and an informative teaching diagnostic – it helps to see it that way. What would you really like to see? You control this, so if you ask them to give you 400 hours of text to read, well, that’s your fault!

What will your visualization look like? What will it do? How will it work? These are important questions for designers and especially learners to ask. Far too often learners, and experienced researchers, create visualisations without thinking what they are doing. Students are often, far to keen to just get coding. When they jump into their code they create solutions that may not be fit for purpose. And when they realize what they have created, it is too close to the deadline to change their mind and adapt it. This early enthusiasm is admirable, and should be encouraged and tapped. Indeed, with some forethought – by becoming more reflective at an early stage, and performing critical thinking – they will create something better. Through thinking and sketching they will be able to contemplate how their solution will work, think who will use it, even imagine a specific person using their tool.

In this lightening talk I presented the need for “visioning”. I proposed that sketching solutions can help people think through their ideas, externalize their thoughts, and project their minds to imagine people using their solution for real, and for its intended purpose.

So to achieve this act of visioning, I proposed that people need to understand their “goal”. The goal then helps to frame the challenge and define the focus. Their goal could be a task they want to fulfill or challenge to solve. In addition, and especially in an education setting, I proposed that these tasks (assessments) should be authentic in their design [1]. In other words, that the tasks should be challenges that they could find in their real life (perhaps when the students have a job after they graduate). Furthermore, the task should be individual to each student.

But learners need structure. They need methods to follow. To scaffold these vision sketches, I proposed a few techniques. The Five Design-Sheets [2, 3] method uses five sheets of paper, with five stages to help drive critical thinking. Alternatively, for specific tasks a single sheet of sketching and planning could be used. One method is the Critical Thinking sheet [4], which gets students to think about the goal, sketch what the solution would look like, list component parts, articulate algorithmic steps, and list tasks that they need to achieve in order to implement it. Figure 8 shows how a student, thinking about a random pattern generator, starts with a sketch that presents the vision of their solution, before implementing and iterating towards their solution in code.

On the first morning of the seminar, I facilitated Sketching Introductions, a 90-minute icebreaker activity that invites people to make personal visualizations through simple drawing. I invited everyone in the group to draw a quick sketch, then introduce themselves using their sketch – people drew with pens or colored pencils on paper to sketch their responses to the question, “How do I see myself in relation to the topic Visualization Empowerment: How to Teach and Learn Data Visualization?”

The group sketched for 7 minutes accompanied by background music, then everyone introduced their sketches in 1-minute introductions. The remote attendees presented their sketches on a shared digital whiteboard via videoconferencing. The in-person attendees projected their sketches to both the in-person and remote attendees. Though some people could not finish their introduction in 1 minute and required more time, everyone completed the activity, and the sketches were as distinct as the personal experiences that they depicted. Each introduction proposed personal approaches to teaching and learning in the VIS context. This icebreaker was a creative visualization activity that enabled group members to share and be introduced to the many perspectives to teaching and learning.

This Vis Futures activity demonstrated a sketching exercise that uses design fiction, collaboration, and creative ideation to encourage players to envision opportunities, use cases, and designs for future visualizations.

The activity uses a set of themed playing card prompts, which emerged from a 2020 workshop on Vis Futures ⁶⁶6https://visfutures.github.io/ at IEEE VIS. Using the Situation Lab’s The Thing From the Future ⁷⁷7http://situationlab.org/project/the-thing-from-the-future/ (a similar sketching game focused on more general future ideation) as a template, the attendees a set of roughly 40 attendees spent several hours proposing, designing, and playtesting a diverse set of different visualization-specific design prompts and cards. Both the original workshop and the resulting game were designed to encourage the use of design futuring to envision the next generation of vis tools and applications.

Since the conclusion of the Vis Futures workshop in 2020, a team of collaborators at the University of Calgary, University of Victoria, and Simon Fraser University have collaborated to develop and refine decks of playable cards to support the activity. These include an online version of the game developed at the University of Victoria ⁸⁸8https://observablehq.com/d/51a981cf418ab2ac as well as a physical card deck currently under production at Calgary.

During the Dagstuhl Seminar, attendees participated in a play test using an in-progress version of the physical cards. In this version, players form teams of 3-5 players. A dealer then composes a sketching prompt by dealing one card from each of four decks – Audience, Data Type, Data Characteristics, Utopia/Dystopia. For example the prompts dealt at the beginning of the Dagstuhl activity included “Student” (Audience), “3D” (Data Type), “Cherry Picked” (Data Characteristics), and “Dystopian Football” (Dystopia). Players then have 5 minutes to independently imagine and sketch possible visualization designs based on the prompts. Afterwards, players share and discuss their designs.

At the end of the short session attendees shared a number of their creative designs and reflections, and offered a variety of constructive suggestions for adapting the gameplay to different audiences and settings. Attendees also voiced considerable enthusiasm for the arrival of the complete, playable card game in Fall 2022.

As information visualization educators and teachers, we are all using a variety of activities to introduce, teach, familiarize students and other audiences with concepts relative to information visualization. Since now we have a poor overview of these types of information visualization activities teacher and workshop organizer are using. We wanted to reflect on these practices with the Dagstuhl Seminar participant in order, to have a better idea of the design space, of what are the meaning full categories to describe this space, what have been deeply explored, what has not been explored, and last what could be generated.

We collected 42 activities from more than 13 different authors gathered from three different sources: 1) the seminar participant and 2) IEEE VIS – VIS activities workshop 2020 [1], 2021 [1] and 3) few other ones we know. On this basis we generate a spreadsheet listing all these activities and created a card deck in which all activities are represented by one card. This card includes (title, author, description, keywords and URL to a document describing it). Each working group received a deck of 42 cards and the link to the spreadsheet.

The activity happened in two main steps 1) design space creation, 2) presentation & discussion. The prompts for the first step were the following “Categorize these activities in a design space from the focus of your group.” It will be in the room for local participants and on a Miro Board ⁹⁹9https://miro.com/app/board/uXjVOp4UJmI=/?share_link_id=161478137131 for remote participants. The prompt for the second step was “Present the design space and the dimensions, maybe the case that was problematic, and the rationale behind each dimensions”. They were asked that it would be totally fine to remove some cards if they do not fit the focus of their group, or even complete the card with activities that were not described in the data set by using a post-it notes or other papers. Last the participant was asked to use the cards as tokens and look for more details in the spreadsheet (images, descriptions, paper, links).

The activity last one hour and 30 minutes, the participant spend 40 minutes to create one design design space by the five groups, and we spent the rest of the time for presenting theses spaces and discussing them. Each group have been able to create meaningfull dimensions.