Robert Kosara, a prominent researcher in the field of data visualization and a former senior staff research scientist at Tableau, delivered a comprehensive presentation at the 2022 Outlier conference that challenged the prevailing preference for standardized, minimalist data representation. The session, which was later released for public consumption on April 26, 2022, was titled The Joys and Dangers of Bespoke and Unusual Chart Types, though it became widely known by its more provocative subtitle, This Should Have Been A Bar Chart! In this address, Kosara explored the delicate balance between the efficiency of traditional charts and the communicative power of custom-designed, or "bespoke," visualizations. The talk serves as a critical examination of how information design can move beyond mere utility to achieve greater audience engagement and memorability.
Chronology of the Presentation and Public Release
The discourse surrounding bespoke visualizations began in early 2022 during the Outlier conference, an annual event organized by the Data Visualization Society (DVS). The DVS, established in 2019, has quickly become a central hub for professionals in data journalism, academic research, and business intelligence. The Outlier conference specifically focuses on the "outliers" of the field—the practitioners and techniques that push the boundaries of how data is perceived and understood.
Kosara’s presentation was first delivered to a live virtual audience in February 2022. During this initial session, the concepts of "extraordinary ways to show ordinary data" resonated strongly with the community, sparking immediate debate on social media and professional forums regarding the "fun" versus "function" dichotomy in professional dashboards. Following the live event, Kosara spent several weeks refining the supplemental materials, culminating in the publication of the full video and a comprehensive reference list on his long-standing visualization blog, EagerEyes, on April 26, 2022. This public release allowed for a broader analysis of his proposed taxonomy and the research-backed theories supporting non-traditional chart types.
The Core Thesis: Functional Efficiency Versus Aesthetic Engagement
At the heart of Kosara’s presentation is a critique of the "Data-Ink Ratio" philosophy popularized by Edward Tufte, which often leads practitioners to strip away any element that does not directly represent data. While this approach minimizes clutter, Kosara argues it can also lead to visualizations that are "devoid of any fun" and ultimately uninteresting to a general audience. The central premise is that while a bar chart may be the most accurate way to compare two values, it is not always the most effective way to capture a viewer’s attention or ensure the information is retained.
Kosara posits that "bespoke" charts—those designed specifically for a unique dataset or context—can serve as powerful tools for storytelling. These charts are often context-specific and may not be applicable to other datasets, but their uniqueness makes them "stick" in the viewer’s mind. By making a chart stand out, the designer encourages the viewer to look closer and engage with the data on a deeper level. This engagement is crucial in an era of information overload, where standard charts are often overlooked as repetitive or "boring."
Research Foundations and the Brehmer-Kosara Collaboration
The theoretical framework for Kosara’s Outlier talk is rooted in academic research, specifically a paper he co-authored with Matt Brehmer, titled "From Jam Session to Recital: Synchronous Communication and Collaboration Around Data in Organizations." This paper, presented at the IEEE VIS conference in 2021, examines the different modes of data communication within professional environments.
In the "Jam Session" phase, data exploration is fluid, messy, and internal, often relying on standard charts for quick analysis. However, in the "Recital" phase—where data is presented to stakeholders, clients, or the public—the need for a polished, impactful delivery becomes paramount. Kosara’s phrase, "to show ordinary data in extraordinary ways," stems from this research. It suggests that once the analytical work is complete, the final presentation should prioritize the narrative and the audience’s emotional or intellectual connection to the data. This transition from "exploratory" to "explanatory" visualization justifies the use of more complex or unusual chart types that might be inefficient for raw analysis but are highly effective for public-facing communication.
A New Taxonomy: Familiarity and Specificity
One of the most significant contributions of Kosara’s talk is the introduction of a taxonomy for evaluating elaborate charts. He proposes that charts be analyzed along two primary axes: chart familiarity and data specificity.
Chart Familiarity
This axis measures how recognizable a chart type is to the average viewer. On one end of the spectrum are high-familiarity charts, such as bar charts, line graphs, and pie charts. These require little to no "learning time" from the audience. On the opposite end are low-familiarity charts, such as Sankey diagrams, chord diagrams, or entirely new geometric constructions. While low-familiarity charts can be visually striking, they carry the "danger" of confusing the audience if not properly annotated or introduced.
Data Specificity
This axis refers to how closely the visual representation is tied to the specific subject matter of the data. For example, a chart about environmental data might use the shapes of leaves or trees to represent values. High specificity makes the chart instantly recognizable in terms of its topic, but it often limits the chart’s utility to that single dataset. Low specificity involves using abstract shapes (dots, bars, squares) that can represent anything from financial earnings to population growth.
Kosara argues that these two axes are independent but related. A designer might choose a low-familiarity chart with high data specificity to create a "bespoke" experience that is both unique and contextually relevant. This framework provides a strategic approach for designers to decide how far they should deviate from standard formats based on their target audience’s literacy and the goals of the communication.
Supporting Data and Industry Implications
The shift toward bespoke visualization is supported by emerging data in the fields of cognitive psychology and digital analytics. A landmark 2013 study by Borkin et al., "What Makes a Visualization Memorable?", found that "visualizations that are highly memorable are often those that are unique, include ‘human recognizable’ objects, and have a higher density of visual features." This aligns with Kosara’s argument that "boring" charts are often the least effective for long-term retention.
Furthermore, data from digital newsrooms such as The New York Times and The Washington Post indicates that "scrollytelling" features—which often utilize bespoke, interactive visualizations—tend to have higher "time-on-page" metrics compared to articles with static, standard charts. These "extraordinary" visuals act as a "hook," drawing the reader into the narrative and encouraging them to explore the data points individually.
However, the "dangers" cited in the talk’s title remain a point of concern for the industry. The primary risk of bespoke charts is the potential for misinterpretation. When a designer prioritizes "fun" or "uniqueness," they may inadvertently obscure the actual data values, leading to "chart junk" that hinders rather than helps understanding. Kosara’s taxonomy serves as a safeguard, encouraging designers to weigh the benefits of engagement against the costs of cognitive load.
Professional Responses and Broader Impact
The reaction to Kosara’s presentation within the data visualization community has been largely positive, though it has sparked a necessary debate about the democratization of data. Critics of bespoke charts often argue that they raise the barrier to entry for data literacy, requiring viewers to learn a new "visual language" for every chart they encounter. Conversely, proponents argue that standard charts have become so ubiquitous that they are frequently ignored, and that "extraordinary" visuals are necessary to maintain the relevance of data in public discourse.
The implications of this shift are particularly relevant for business intelligence and corporate communication. As organizations move away from static reports toward dynamic storytelling, the ability to create "bespoke" visuals that align with a brand’s identity or a specific project’s goals is becoming a highly valued skill. Kosara’s talk provides a roadmap for this evolution, suggesting that the future of the field lies not in the total abandonment of the bar chart, but in the strategic use of more creative, context-aware visual forms.
In conclusion, Robert Kosara’s The Joys and Dangers of Bespoke and Unusual Chart Types serves as a pivotal call to action for the data visualization community. By advocating for "extraordinary ways to show ordinary data," Kosara encourages practitioners to embrace the complexities of human perception and the power of aesthetic engagement. While the bar chart remains a vital tool in the analyst’s toolkit, the bespoke chart is increasingly recognized as the essential tool for the communicator, bridging the gap between raw data and meaningful human insight.