Feeling Numbers: the Rhetoric of Pathos in Visualization
Total Page：16
File Type：pdf, Size：1020Kb
Feeling Numbers The Rhetoric of Pathos in Visualization SARAH CAMPBELL Feeling Numbers The Rhetoric of Pathos in Visualization Thesis presented by Sarah Campbell To The Department of Art + Design In Partial Fulfillment of The Requirements for The Degree of Master of Fine Arts in Information Design + Visualization Advisor: Dietmar Offenhuber Reader: Alberto Cairo Respondent: Pedro Cruz Chair: Nathan Felde Northeastern University Boston, Massachusetts May, 2018 Acknowledgements To my mom and dad, for their never-ending support. To Alex, for being my rock through this journey. To my cohort, for the critiques, laughs, shared pain, and inspiration. To Phelps, for being such an entertaining turtle. Abstract Rhetoric is a powerful tool used to influence and persuade. Due to their inherent subjectivity, data visualizations are a form of communication that employ persuasion techniques. Therefore, the rhetoric of visualizations deserves deeper investigation. Drawing from Aristotle’s three modes of persuasion, this thesis explores the rhetoric of pathos, or appeals to emotion, within data visualizations. In this thesis, I develop a taxonomy of pathos techniques applied to visualizations and empirically measure the emotional effect of pathos techniques that relate to data proximity. This research improves the visualization community’s understanding on how certain design decisions can add meaning and relevance to data. Contents 11 17 37 Introduction Rhetoric and Techniques for Emotional Appeals Appealing to Emotion: in Visualization A Taxonomy Rhetoric Engage Emotional Appeals Humanize in Visualization Personalize 49 73 77/83 Evaluating the Conclusion References Emotional Effect of Proximity Techniques Appendix The Value Proximity Emotion Study Methodology Analysis & Results Discussion Introduction Data and Emotions. These two words are an unlikely pair. Data are typically described as rational, objective, cold, black and white. Emotions, on the other hand, are subjective, colorful, and expressive. It may seem like the two do not have a common meeting place, but they potentially do: in the visual representation of the data. Today data are ubiquitous, so it is not surprising that visualizations of data are increasing in ubiquity as well. With that, visualizations have broadened their scope to act as a communication tool to convey powerful messages. Data journalism has popularized visualizing information for the general audience. Newspapers like The New York Times, The Guardian, and The Wall Street Journal employ their own graphics teams that visually convey information on issues we care about: politics, the environment, health care, education, and so on. Advocacy groups are also taking advantage of increased accessibility of data, harnessing the power of information and technology for activism by appealing to audiences rationally, ethically, or emotionally (Tactical Technology Collective 2014). Appeals to logic, ethics, and emotion lie at the foundation of Aristotle’s Figure 1.1 system of rhetoric (2004). The existence of persuasion in visual Items: Is Fashion Modern? A representations of data is not consensually accepted in the visualization fashion exhibition visualized community. When persuasion is discussed, it is typically discussed by Accurat for MoMa through negatively connoted terminology such as bias (Hullman and (Bassan et al. 2017). 12 INTRODUCTION Diakopoulos 2011) or deception (Pandey et al. 2015). The study of rhetoric in this way has tremendous value for developing a visual literacy to recognize forms of deception, but I argue that rhetoric has other facets to consider with respect to data visualization. If we look at persuasion as an umbrella term for influence, whether it is to influence one’s beliefs, attitudes, or behaviors, the concept is not completely negative. However, efforts to influence by visualizing data have been traditionally looked down upon due to visualizations being considered as objective, neutral representations of the data. This viewpoint is fueled by the scientific, analytical lens that visualization has typically been looked at through. Additionally, influence in visualizations is typically focused on in terms of unethical design decisions that lead to deceptive representations of the data, where the message of the data does not match the message of the visualization. In an effort to minimize deceptive uses of data, the visualization community has emphasized data-centric designs, where clarity of the data is stressed. Therefore, any visual elements that do not aid in maximizing the data shown are considered unnecessary. However, in several ways, this has led to false thinking that there is an objective way to visualize data, when indeed there is not (Dörk et al. 2013; Hullman and Diakopoulos 2011; Viegas and Wattenberg 2007). If this were true, designers given the same data would visualize it in the exact same way, making identical design decisions. In reality, one dataset can be visualized in many different ways. By merely going through the process of visualizing data, the designer is creating a persuasive message, since the act indicates that the designer believes the data can be influential in some way. In this light, it is far more ethical to acknowledge the subjective nature visualizations have instead of masking it under the ideals of neutrality. Visualization practitioners first need to acknowledge the subjective, persuasive nature of visualizations and look at the practice as a form of communication. By better understanding the subjective and influential aspects of a visualization, we can better scrutinize design decisions made that deviate from the message of the data. We can go a step beyond that and embrace the subjective nature of visualizations. In doing so, we can craft visualizations that are more human-centric for our audience: people. A closer look at persuasive efforts in this way focuses on pathos: the appeal to emotions of the audience. In Aristotle’s three established appeals, pathos is the mode that is the least agreed upon in the visualization community. It is understandable how an accord was found in the use of visualizations to speak to reason and ethics. But to admit, or even celebrate, the potential for visualizations to bring an emotional capacity to data? This is a concept the community as a whole has only recently begun to talk about. With the growth of using visualizations INTRODUCTION 13 as a form of communication about important issues relevant to humanity, designers are seeking ways to add the human element to the data and impact how the audience feels about the information the visualization conveys. Through recent literature, conference talks, blog posts, and podcasts, practitioners are engaging in a conversation about emotions in visualizations. Some even offer techniques that may appeal to the emotions of the audience. However, little research has evaluated the emotive impact these techniques have in visualizations. Do they evoke emotion at all? If so, how powerful are they? This thesis builds upon existing discourse and research on the role of pathos in visualizations. Through collecting and categorizing recommended techniques, I propose a working taxonomy of techniques that are said to appeal to emotions. More importantly, I present an empirical evaluation of three specific techniques: proximity of interests, temporal proximity, and spatial proximity. These techniques rise from the concept that we care more about people and events that are near us in time, space, and importance (Campbell 1776). Through this evaluation, I begin to identify the emotive impact these techniques have when implemented in visualizations. The goal of this work is to further the research on how visualizing data can appeal to emotions. By improving the community’s understanding of emotional appeals, we improve our understanding on the communicative power of visualizations as a whole and, therefore, enhance our ability to visualize data to connect with the people who explore it. 14 INTRODUCTION INTRODUCTION 15 If I look at the mass I will never act. If I look at the one, I will. Mother Teresa Rhetoric and Emotional Appeals in Visualization This chapter serves to highlight related work around rhetoric and emotional appeals in visualization. While rhetoric has established theories throughout history, the visualization community as a whole has yet to cozy up to the idea of visualization as rhetorical communication. However, visualization literature points to the subjective and rhetorical nature of visualizations. When it comes to appealing to emotions, psychology research validates the struggle to affect people with numbers. Data visualization practitioners acknowledge this difficulty and, at the same time, many see value in arousing the audience’s emotions through visualizing data. A few practitioners identify techniques that are suspected to appeal to emotions in visualization, but very little research evaluates the emotional impact of these techniques. RHETORIC Aristotle is credited with developing the foundational system of rhetoric and Figure 2.1 defines it as the study of means of persuasion (2004). Within this system are Off the Staff. a series three modes of persuasion. The first is logos, which appeals to reason. Since of scores visualized by rhetorical efforts were primarily delivered in the form of speeches in his time, Nicholas Rougeux for Aristotle refers to logos as the substance of the speech itself “when we have the OpenScore initiative proved a truth or an apparent truth”