Affective Design For Data Stories

Data stories are data-driven visual mediums that integrate storytelling with data visualization design.Since data stories mainly serve the goal of data communication,in recent years,researchers have increasingly paid attention to users’ experience with data stories,and affects(i.e.,emotions),as a key indicator of user experience,have also been emphasized.However,although many researchers have recognized the importance of affects,so far,relevant literature is very limited,and little work has been done to thoroughly investigate the role of affects in data stories.Given such motivations,this work aims to examine how the design factors in data stories relate to affects from various perspectives and construct a set of design spaces that describe the spectrum of affective design methods for data stories.The structure of this dissertation is as follows.First,Chapter 3 explains the pipeline of creating data stories,based on which a “five-element” model for affective design in data stories is proposed.Then,by referring to the model,this dissertation explores the relationship between design and affects in data stories from multiple perspectives.Chapter 4focuses on how the graphical elements in infographics help convey affects.Through two user studies,we identified 12 common affects evoked by infographics and developed a structured framework of "design heuristics" to guide the design of affective infographics.Then,we attempted to quantify the relationship between graphical elements and affects more precisely through the lens of affective arousal(Chapter 5).We first conducted a crowdsourcing study to collect user ratings for data visualization designs,then extracted design features from the images,and then constructed a model to predict the relationship between these features and affective arousal.As a result,we identified 13 design features that had significantly influenced affective arousal.In Chapter 6,we investigated how narrative order affects affective engagement.We identified six common narrative orders in time-oriented stories,and our evaluation study showed that several non-linear narrative orders performed significantly better than chronology in terms of affective engagement without hindering the comprehensibility of data.In Chapter 7,we investigated how to convey affects with animated charts.After the iterative design,we proposed a design scheme for affective animated charts and validated the effectiveness of these designs using two user studies.Finally,we examined serious data stories that convey negative affects and studied multiple design elements(e.g.,graphical elements,animation,interaction)together(Chapter 8).Through two user studies,we identified 19 design techniques that may help augment negative affects in data stories and found that negative affects had significantly enhanced the participants’ contemplative experiences.Chapter 9 summarizes the findings of this dissertation and envisions possible future research directions and application areas.Through the above research,this paper pioneers the research around affects in data stories by systematically examining affective design in data stories for the first time.Besides,affective design,as a cutting-edge research paradigm,can play an important complementary role to the traditional data visualization research paradigm,assisting the design or the automatic generation of affectively expressive data stories in the future.