In an increasingly polarized world, even well designed charts were sometimes twisted to make a case against public health advice<\/a>. Where we sourced our charts shaped the stories curated from the same datasets<\/a>, a stark reminder that \u201cnothing about a chart is inherently objective.<\/a>” Chart annotations helped to add context, titles read as headlines instead of lists of metrics, and graphic legends helped more novice chart readers focus on the key messages in charts. Yet, wide gaps in how people see the same numbers and understood them to show entirely different patterns underscored gaps in data literacy and specifically data visualization literacy, even in a world where we are inundated with charts. <\/p>\n\n\n\n The role of the data visualization creator is more critical today than ever before. It is no longer enough to make a good chart. Published charts must be readable and durable across audiences and contexts in a world where screenshots shared on social media can make a chart go viral for all the wrong reasons. If we make charts for public sharing that require us, as creators, to teach people how to read them, have we really made information more accessible? Or just put something out into the world to be ignored or, worse, misused.<\/p>\n\n\n\n Until we grow the data visualization literacy broadly, then reputation and affiliation will always be more important than the objective quality of the chart. Think simply about how we perceive charts from different news sources or content creators, particularly those who have a distinct political agenda. The simple efficiency of encoding visual variables have defined data visualization since its popularization in The Grammar of Graphics: communicating ideas through shape, color, size, and more. The world today still demands charts that communicate clearly (with attention spans that feel like they\u2019re ever-shrinking), but as a field we need to embrace the messy complexity of charts being used as memes, political ads, and protest. This means expanding into more complex design components, visual metaphor, and data exploration.<\/p>\n\n\n\n As we mark a shift into an era of AI enablement, grapple with pushes towards greater polarization in our politics, and see how charts shape public sentiment on key issues like health and climate, access to information is more important than ever. In a world where we are saturated with content and have greater access than ever before, how can the field of data visualization take a leading role in ensuring we\u2019re facilitating the use of that information?<\/p>\n\n\n\n In this article, we revisit the waves of clarity, systems, and convergence defining the era of modern, computer-enabled data visualization and make the case that we\u2019re shifting into a new fourth wave: the democratization of data visualization, bringing both opportunities and challenges to be realized by data visualization creators.<\/p>\n\n\n\n Advances in our digital tools and the use of data visualization in society has been deeply informed by charts made of physical objects and hand illustrated graphics: a history of data visualization would feel woefully incomplete if the timeline started in the 1950s and excluded names like Nightingale, Minard, Playfair, Snow, and others. Go back even further and we find evidence of encoding information in knots, colors, tapestries, and more dating back thousands of years.<\/p>\n\n\n\n But the rapid speed of change that came with advances in computer-based charting is something different. In the keynote of the Tapestry Conference in 2018, we conceptualized three waves of data visualization<\/a> in practice during the era of digital design. The first wave focused on clarity, the second on systems, and the third on convergence within the field.<\/p>\n\n\n\n The standards from the emphasis on clarity of the first wave pushed for decluttered charts. The concept of removing \u2018chart junk\u2019 and being mindful of the \u2018data-to-ink\u2019 ratio popularized by Edward Tufte in his original book defined the objective of visualization as moving charts towards simplicity, without space for art or embellishment. Other thought leaders accelerated the push for improving the function of visualization through<\/em> form, advocating for actionable headlines and thoughtful annotations. Those core principles are cemented into the practice of data visualization and still felt today. Research has validated some of these principles, like how a declutter and focus approach to chart design makes the key takeaways in charts more memorable<\/a>.<\/p>\n\n\n\n The second wave, systems, brought shared language and libraries that enabled dramatic new works focused on encoding data to channels of color, shape, position, and anything else you could draw. This fostered more collaboration and conversation between data visualization creators focused as much on creating new forms as on celebrating that creation. But it also created tool specific silos with D3 developers, R coders, Tableau designers, and other stack-specific communities carving out their own approaches, experts and best practices.<\/p>\n\n\n\n The third wave, convergence, brought together these shared principles and conversations, with what we create less defined by the tech stack we\u2019re using. Analysts, coders, designers, journalists, and data artists connected through community spaces and social media, sparking cross pollination across domains and tech stacks that ignited renewed excitement for was was possible<\/em> through data visualization. Tableau users began creating more infographic-like and creative visualizations rather than just using the tool for exploratory analysis and business dashboards; notebooks and Shiny apps allowed coders to create friendly scrollytelling stories and dashboards with tools often assumed to produce singular graphics to embed in a website or other space.<\/p>\n\n\n\n Through the continued advancements and exploration of human cognition research, we\u2019ve also built a large body of knowledge on what works when creating explanatory graphics<\/a>. And yet we still have spaces to explore, particularly around the understanding and interpretation of charts beyond the typical research subjects recruited by US or European studies.<\/p>\n\n\n\n This convergence of different tools, principles, and ideas to create visualizations for a wide array of audiences and purposes was furthered with the founding of the Data Visualization Society<\/a>, the global professional association for data visualization practitioners and enthusiasts. The seeds of DVS were sown at the Tapestry Conference with the closing call to action for more collaboration across tool silos, and the official organization\u2019s founding came a few months later in February 2019. <\/p>\n\n\n\n Today, six years later, engagement with data visualization at work and, more importantly, in our lives, has grown. The wide use of charts in our day-to-day lives (from news articles to bank statements to Apple Watch rings) means we don\u2019t even register that we\u2019re engaging with data sometimes. This ubiquity of charts, coupled with the emergence and rapid evaluation of AI technology, is reshaping how we engage with information, and, in turn, the field of data visualization.<\/p>\n\n\n\n In some ways, generative AI is just another technological change to the way we produce data visualization, like introducing charting libraries in Excel or the Flare library in Java or D3. GenAI promises an even greater democratization of the means to create data visualization. But does the human sit in the driver\u2019s seat or are they relegated to just being a part of the feedback loop?<\/p>\n\n\n\n Tools that used to require deep technical knowledge like matplotlib or D3 are suddenly accessible. Because ChatGPT is better trained on code than using existing user interfaces, at the time of writing this paper the platform perhaps ironically makes it easier for a novice to make a chart with code than with a low-code BI tools. Not that BI tools have missed this\u2014if you haven\u2019t seen GenAI integrated into your favorite charting tool, just wait, it\u2019s coming. And if you have (which you surely have by now) it\u2019s only going to increase in its penetration.<\/p>\n\n\n\n GenAI is yet another technology that has simplified the creation of basic charts, but complex graphics still remain the domain of viz experts with many requiring programming skill or artistic capabilities to create. But the future is clear: technical skill will fade as the gatekeeper of complex data visualization. This is a good thing: we need more people invested in the craft and creation of visuals that help us understand a complex world.<\/p>\n\n\n But more urgently, we\u2019ll need to address the data literacy and graphicacy gaps that have long been a challenge in democratizing engagement with data. Where people are using genAI tools, they may find welcome companions in decoding the messages in a chart. Where AI tools can ingest multimedia prompts, you can ask for a read out of the key messages from a graphic. But this continues to rely on access to technology, risks hallucinations or simple mistakes, and raises significant concerns around the explosive demand for energy to fuel all of these queries.<\/p>\n\n\n\n We need to get ahead of this growing creation and dissemination of more complex forms by emphasizing the role of data visualization professionals in fostering data visualization literacy. We need to focus on readers\u2014and on reading charts\u2014whether that reading happens via or alongside AI.<\/p>\n\n\n\n As we move into 2025, genAI tools can read simple charts, but don\u2019t excel at interpreting complex data graphics. These tools perform a bit better at gleaning insights when we share some code or data signature for the tools to reference, which adds complexity and labor for whomever is doing the prompting. Does that mean the use of AI tools will drive us toward even simpler chart types? In the short term, probably, but like everything else in genAI it will likely mature faster than we expect.<\/p>\n\n\n\n But what about more complex interactives and longer data stories? When expert designers have experimented with genAI to create visual data stories, as illustrated with a very practical set of prompts and a report card on ClaudeAI\u2019s performance by The Pudding, the new technology can be powerful collaborators in the data cleaning and synthesis stages, but don\u2019t make passing grades for creating engaging final products<\/a>.<\/p>\n\n\n\n Observable itself transitioned from enabling bespoke weird graphics to promoting Observable Plot, a traditional charting library. Tools like Observable also deliver mixed code and interface capabilities as a matter of course. But these tools still don\u2019t know how to mix code, user interface design, and AI outputs together into a cohesive final product, though that future looms large with work being done in this area by companies like HEX<\/a>.<\/p>\n\n\n\n This will change the field as much as graphic user interface driven tools and modern data visualization libraries like D3, ggplot2 and matplotlib did in years past. Just as it is no longer enough to be able to make charts or create geometrically sophisticated experiments, it will not be enough to make good charts. We need to make good readers. When data visualization becomes cheap and easy to make, its value will be in how we use it, recognizing it as a functional, artistic, exploratory, analytical, and meaningful practice.<\/p>\n\n\n\n The changes in the field bring us to a fourth wave of data visualization, one defined by the ways data visualization plays a critical role in shaping how all of us<\/em> navigate our world. With the ubiquity of charts and data in social media, news, mobile applications, work, and more, having a basic knowledge of how to read charts and graphs is more important than ever before. The fourth wave of visualization is defined by its democratization of the creation and use of charts to inform, inspire, and shape decisions (usually for good, but also sometimes for ill).<\/p>\n\n\n\n Data visualization creators play a critical role in ensuring the meaning, usability, and trustworthiness of graphics. Who created the chart, including if AI played a supporting or leading role, will be an even more important question in this fourth wave as charts shape our understanding of the world. Whether that creator\u2014person or machine\u2014understood the principles of data design and used them ethically is the first question every reader should ask when confronted with a chart. Then, what insight is the chart conveying and is it backed up by the data\u2014or is it cherry picked to make a point. These are questions many will need to learn to ask<\/em> as we help others think critically about each chart they see.<\/p>\n\n\n\n Conversations about the evolution of data-related fields often center on technology. But with the role data visualization plays bridging from data tables to usable information, we must center people\u2014both the readers and creators of visualizations\u2014in the ways we conceptualize the design and use of charts in this fourth wave.<\/p>\n\n\n\n How can we make data visualization accessible to more aspiring designers and readers alike? We need to ensure we\u2019re speaking a shared language of data visualization, and considering our audience\u2019s needs first in the design process. Meet people where they are, and then bring them along to understand increasingly complex graphics. We know more people can make charts today than ever before, but do they have the theoretical grounding to design them effectively? Do we need to teach people how to make charts in order to make them better readers of charts?<\/p>\n\n\n\n Five necessary tenets define this fourth wave and the continued democratization of data visualization: speaking a shared language, embracing non-traditional forms, engendering trust, embracing complexity, and contextual literacy and ethics.<\/p>\n\n\n\nWhere we came from: clarity, systems, and convergence<\/h2>\n\n\n\n
The rise of AI<\/h2>\n\n\n\n
![\"A](\"https:\/\/i0.wp.com\/dvsnightingale.wpenginepowered.com\/wp-content\/uploads\/2025\/02\/image6-1024x573.png?resize=720%2C403&ssl=1\")
Where we are now<\/h2>\n\n\n\n
The fourth wave: democratization<\/h2>\n\n\n\n