Visual Complexity website. © Manuel Lima.
IG: Trained as an industrial designer, you are currently a senior user experience design lead at Microsoft and a leading voice on information visualization. Can you talk a little bit about your path to your current work?
ML: Even though my background is industrial design I was always interested in the multi-disciplinary nature of design and its various practices, such as graphic design, interface design, motion design, or 3D animation. After graduating I did an internship at Kontrapunkt, a small design agency based in Copenhagen, Denmark, which despite its size was able to produce work in numerous design fronts, including print, web, and product. This was a very enriching experience for me, particularly due to the well-known Scandinavian design charisma. After this internship I applied for an MFA program at Parsons School of Design. This was by far the most important stage of my professional career, where I fell in love with data visualization and information architecture. Since then I’ve become a blogger, researcher, speaker, and author on the topic of network visualization, and I’ve worked as an interaction/UX designer in companies like R/GA, Nokia, and Microsoft.
IG: In October of 2005, you founded Visual Complexity. What is the idea behind it? What are the criteria in the selection?
ML: I started VisualComplexity.com in 2005, in the aftermath of my MFA thesis research, primarily as a personal bookmarking mechanism, to keep track of various topics I was interested in. Later it evolved into a public resource for anyone interested in information visualization, particularly in the mapping of networks. One of the main reasons why it has been quite popular is certainly the breath of content covered in the site. Even though it has a manifested emphasis on networks, it covers any depiction of this engaging topology, from protein networks, subway systems, or social networks.
The key criteria is that a project should either provide advancement in terms of visual depiction techniques/methods, or show conceptual uniqueness and originality in the choice of a subject. In the end, it’s always a personal and subjective decision that strongly takes into consideration the narrow scope of network visualization.
AC: Describe complexity and why a holistic picture is worth drawing?
ML: Complexity normally describes a system with many interconnected and interdependent parts that form an intricate arrangement.
A few days ago I was part of a panel at Ars Electronica in Linz, Austria, and the main topic of the panel, as well as the entire symposium, was the “Big Picture.” The notion of providing the big picture, the holistic picture, or simply, the overview, has never been so critical as nowadays. Even though its implications are vastly diverse, visualization can play a key role in uncovering the various principles and behaviors that regulate many natural and artificial systems, from our own brain and the vast biosphere, to transportation infrastructures and the World Wide Web. Normally characterized by a multitude of interconnecting and interdependent elements, where no unit can be altered without modifying the whole, complex systems present us with one of the hardest scientific challenges of our time. As new methods of analysis, modeling, and exploration are needed, in order to unravel the complex connectedness of modern society, visualization will hopefully roll up its sleeves and answer this vital call.
AC: When should we create “maps” of complex information or “tracings”?
ML: Whenever a given system, structure, behavior, or process is not entirely understood, or new types of knowledge and insight can be ascertained. Cartography, as well as its most recent cousin, data visualization, has always been driven by human curiosity and our desire to know more. Both fields simply aim to provide us with tools for insight, understanding, and sense making.
IG: What have been the changes in data visualization that you have noticed since you started Visual Complexity?
ML: Almost as staggering as the assortment of portrayed subjects is the variety of employed visual techniques. Frequently generated by computer algorithms and enhanced by interactive features, most projects showcase a broad palette of visual elements and variations that consider color, text, imagery, size, shape, contrast, transparency, position, orientation, layout, and configuration. Despite this rich graphical diversity, many projects tend to follow noticeable trends and common principles, which in turn result in a type of emergent taxonomy. This embryonic and evolving taxonomy provides a portrait of the current state of the practice and reveals the initial building blocks shaping a new visual language. This alphabet is not entirely new in the sense that many of its letters are recurrent visual metaphors used for centuries. But some are combining and recombining old metaphors in new original ways. This emergent grammar of visualization, particularly network visualization, has been featured in a dedicated chapter of my latest book, Visual Complexity: Mapping Patterns of Information.
AC: What are three principles that successful visual information graphics contain?
ML: Easily graspable, evermore insightful, conducive to exploration.
AC: What new types of visualization methods have emerged recently? What can we look forward to?
ML: It will be interesting to observe how Information Visualization embraces other emerging technologies, in areas such as Interaction Design (HCI), Physical Computing or Pervasive Computing, and produces interactive experiences that go well beyond the computer screen. Looking at ambient and immersive visualizations can provide us an interesting look into what the future holds for us. I like to envision a time where large amounts of useful information will be less intrusive and more dissolved into everyday objects and surfaces, and ultimately more in tune with human behavior.
IG: Where are you seeing the biggest trend in data visualization being implemented and having an impact?
ML: I would like to think in the awareness of global warming and the impact of human behavior on our planet.
AC: How do you balance information that is simple to understand and yet visually arresting to pull someone into the information?
ML: Some people tend to see a clear, impenetrable divide between function and aesthetics. However, as numerous studies have shown, aesthetics and novelty are a functional feature of most projects, making the user experience more satisfactory, engaging, and ultimately, more memorable. In fact, aesthetics designs are normally perceived as easier to use than less-aesthetic designs, so there’s a clear correlation between aesthetics and usability.
IG: One of the goals, if not the main goal, of data visualization is to turn raw data into meaningful knowledge. Can you give us some successful examples in which information visualization helps to understand complex situations and generate meaningful knowledge?
ML: One of the most well known cases, and in part responsible for the tipping point of the field, from its academic womb to the general public, is the Map of the Market by Martin Wattenberg.1 In this interactive map, a recent classic of the field, one can easily perceive at a glance the daily fluctuations, and inherent complexities of the various sectors of the stock market. Another less-known example is the Cod Food Web, created by David Lavigne.2 This convoluted map, illustrating close to 100 interdependent species in the Northwest Atlantic, was created in order to explain the causes of the cod stock depletion and the interconnected arrangement of its natural ecosystem.
AC: Facebook is the one of the most visualized networks, and yet most visualizations are simply representations of the network. How can a visualization tell us little about the power of a network rather than just the image of the network?
ML: It really depends on the initial question one might ask, or the specific topic of analysis. There are hundreds of ways one could examine, investigate, and display a given network, by concentrating for instance on its usage patterns, over-time interaction, shared content, geographic activity, or users’ typology and demographics. A visualization of its topology is just one of them. Perhaps we’re still in an early exploratory phase, where the focus is more on the disclosure of the structure itself, rather than what’s happening inside it.
IG: In your presentation at the See Conference in Wiesbaden, you describe the evolution of data visualization, from the problems of simplicity of the seventeenth, eighteenth, and nineteenth centuries, to the problems of disorganized complexity of the first half of the twentieth century, and finally to the organized complexity of the second half of the twentieth century and beginning of the 21st century. Can you talk a little bit about that transition? Why is it important to visualize the complex networks that exist between seemingly unrelated things? What is/are the biggest challenge/s we are facing?
ML: In 1948, in an article entitled “Science and Complexity,”3 Weaver divided the history of modern science into three distinct stages: The first period, covering most of the seventeenth, eighteenth, and nineteenth centuries, encapsulated what he denominated as “problems of simplicity.” Most scientists during this period were fundamentally trying to understand the influence of one variable over another. The second phase, taking place during the first half of the twentieth century, involved “problems of disorganized complexity.” This was a period of time when researchers started conceiving systems with a substantial number of variables, but the way many of these variables interacted was thought to be random and sometimes chaotic. The last stage defined by Weaver, initiated in the second half of the twentieth century and continuing to this day, is critically shaped by “problems of organized complexity.” Not only have we recognized the presence of exceedingly complex systems, with an outstanding number of variables, but we have also recognized the notion that these variables are highly interconnected and interdependent.
Many of our contemporary hurdles, from the way we organize our cities to the way we decode our brain, concern problems of organized complexity that cannot be portrayed, analyzed, or understood by employing a centralized or reductionist model. These new complex challenges deal primarily with rhizomatic properties such as decentralization, emergence, mutability, nonlinearity, and ultimately, diversity. Therefore, they demand a new way of thinking. They demand a pluralistic conception of the world, able to envision the wider structural plan and at the same time examine the intricate mesh of connections amongst its smallest elements. They ultimately call for a holistic systems approach; they call for network thinking.
IG: In the same lecture, you point out a few, really interesting examples of information visualization presented as 3D installations, such as Tomas Saraceno’s “Galaxies forming along filaments like droplets along the strand of a spider’s web” (2008) and Chiharu Shiota’s “In Silence” (2008). Are we at the beginning of the 3D exploration in data visualization? What is the potential in 3D versus 2D visualizations?
ML: Many of the “fake” 3D projects (displayed in a 2D screen) produced in the 1990’s have been considerably unsuccessful, primarily because we have a hard time orienting ourselves through data presented in intricate 3D virtual structures and constructs. Instead, real 3D can offer us a stimulating path to explore, and we have witness interesting attempts on the parallel field of generative art, but it still seems like large, highly immersive and multi-sensorial environments can be a more reliable and efficient alternative. The AlloSphere Research Facility, at the University of California, Santa Barbara, is a great example of this new paradigm.
IG: Most of the data visualizations we are seeing right now are representations of data already collected, static information. With all the tracking devices available nowadays, both personal and collective, are we ready to map the dynamics of real time information? What are the challenges? Who is the possible audience?
ML: Time is one of the hardest variables to map in any complex system. It is also one of the richest. If we consider a social network, we can quickly realize that a snapshot in time can only tell us a bit of information about that community. If time were to be properly measured and mapped, it would provide us with a comprehensive understanding of the social group’s changing dynamics, how it expands or shrinks, how relationships evolve, and how certain nodes become more or less prominent. In some cases, the changes do not take weeks or months, but minutes or hours. And it is not only the network that adapts; whatever is being exchanged within the system also fluctuates over time (e.g., information, energy, water, a virus). There is no doubt that when we embrace time, the difficulty of the task at hand increases tenfold, but we need to make this substantial jump. Most networked systems are affected by the natural progression of time, and their depiction is never complete unless this critical dimension becomes part of the equation.
AC: What visualization recently caught your eye?
ML: In Ars Electronica, last week, I was part of a panel with Golan Levin, Adam Bly, Johan Bollen, and Paola Antonelli. During his talk, Golan revisited The Secret Lives of Numbers, a project produced 10 years ago that aims at mapping the popularity of numbers in the Google search engine. An early classic of data visualization, The Secret Lives of Numbers is also a critical example on how curiosity is the main driver for visualization.