The Democratization of Data Visualization

Over the last few months, Elijah Meeks and I have been working on a white paper about the Fourth Wave of data visualization. We felt that the field had moved into a new era, beyond the first three waves he described in his Tapestry Conference keynote in 2018.

The first three waves describe overlapping shifts in the world of modern (mainly computer generated) data visualization:

  • Wave 1: Clarity, where chart creators shift towards decluttered, simplified graphics with clear headlines thanks to the influence of data viz practitioners like Edward Tufte, Stephen Few, Cole Nussbaumer Knaflic, Stephanie Evergreen, and others.

  • Wave 2: Systems, where we focus on organizing our thinking and language around how we describe charts into more standardized frameworks (like the Grammar of Graphics) and libraries (like D3).

  • Wave 3: Convergence, where we see tools and communities focused on creating data visualizations overlap and engage more, with tools being adapted to new purposes. Think of Tableau being used to create data art instead of just business dashboards or analytical notebooks used to create interactive reports.

The visual metaphor of a wave is deliberate: the effects of Wave 1 are still rippling through our field today, rather than being staged phases we move through and abandon.

What came next

The Data Visualization Society was born out of conversations at Tapestry 2018 and Elijah’s call to action that we find ways in the third wave to collaborate, share, and create more deliberate convergences across our industry.

Looking at the backgrounds across our six founding board members, with graphic design, complex coding, consulting, and public health research communication, we were a team keen to create the spaces for conversation and collaboration beyond any one tech stack. We wanted to demystify the paths into careers in data visualization, which were a bit meandering for each of us. Those intentions shaped what we created within DVS, including Outlier and Nightingale.

Six years later, we’ve created a community that brings data viz people together across tech stacks and domains. We’ve also transitioned leadership to other brilliant data viz practitioners who will set the course for the next five years and beyond.

In that same time, our field has also evolved. COVID placed charts on the front pages of every major newspaper. We collectively experienced a pandemic through the language of charts, making us more familiar with more complex chart types - even log scales.

Then, GenAI started reshaping how we seek and communicate information in a big way. The share of data viz creators responding to our State of the Data Viz Industry survey who said they used AI in their work jumped from 24% (2023) to 37% (2024).l Anecdotally, I see more chart creators talking about how they use GenAI tools in their workflow.

Bar chart and upset plot showing where respondents used AI tools in their data viz workflow in 2024 (Data Viz Society SOTI Report 2024)

Tools like Claude and ChatGPT have evolved to scrape, analyze, and visualize data, reducing some of the remaining technological barriers to creating charts by allowing people to use natural language prompts to build a graphic.

But the evolution of this technology also raises pressing questions about who creates charts and how readers assess trustworthiness

What ‘democratization’ means to us

With each of the first three waves defined by a word—clarity, systems, and convergence—we mulled on what word would define this new wave. Data visualization has transformed from a niche, specialized tool into a fundamental part of how we understand politics, public health, and culture.

We’re now in a fourth wave defined by the widespread ability to make and share visualizations: the democratization of data visualization.

I’ve had many conversations around this democratization enabled by technology as the landscape of data viz design tools grew over the last decade, often in the context of career conversations around ‘what tools should I learn?’. Tools like Datawrapper, Flourish, Observable Plot, RAWgraphs, Canva, and more make spinning up a chart without any code easier than ever before, with design defaults that can enable anyone to make a decent looking graphic.

But when data visualization becomes cheap and easy to make, its value will be in how we use it, recognizing viz as a functional, artistic, exploratory, analytical, and meaningful practice. We’ve seen the expansion of creative, non-traditional forms of expression, including hand-drawn and physical installations with roots that go back centuries to the days when all data visualization was done by hand.

The hyperwall at the NASA Earth Information Center in Washington D.C. showcases a 20 minute loop of charts, maps, data stories, and purely aesthetic visualizations that help viewers learn about the changing climate. (Photo by Amanda Makulec, all rights reserved)

As data visualization becomes more accessible, it also becomes more susceptible to misuse. Quantity doesn’t necessarily come with quality information. Which brings me to the question: is the democratization of data viz a good thing?

While you don’t have to read the full white paper, I hope you will particularly if you’re working in data viz today. We unpack more details around five key tenets that underpin this fourth wave: 1) speaking a shared language, 2) embracing non-traditional forms, 3) engendering trust, 4) embracing complexity, and 5) contextual literacy and ethics.

Is this shift good for data viz creators?

When we ask if a change is good, we also have to ask ‘good for whom?’. First, let’s tackle if this democratization of data viz is good for those of us who create data visualizations.

Yes, because our field will grow.

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. We can invite people who are from data-adjacent fields to learn the art and science of visualization design. Their subject matter expertise adds value and strengthens the diversity in our field, as many of us have by bringing our not-analytics-degrees to this space (my own credentials are in zoology, sociology, and public health, so perhaps I’m biased there).

Reducing gatekeeping doesn’t mean we remove the need to invest in learning though. Creating meaningful charts still requires taking care to learn the fundamentals of data viz. Removing barriers means simple, straightforward charts are increasingly easy to create if you can work with a data table in one of the many new tools, and fine tune a chart with some design knowledge. These are skills that can be learned.

As established data visualization practitioners, we have an opportunity to step up to share our knowledge with those new to the field, whether through formal mentorship, established communities of practice, or other forums. We can help new practitioners see beyond promises from tech tools and think more deeply about how their design decisions shape how an audience sees a chart.

Yes, because tedious parts of our work can be made more efficient.

For data viz designers, we have new tools to make our work more efficient. We can debate the impact of these tools on the creative process of building charts, but they aren’t likely to wholly replace data viz creators soon.

When used thoughtfully, emerging tools and AI integrations within our existing charting platforms can help us automate data scraping, cleaning, organization, and even crafting documentation that many of us find tedious, as long as we abide by the mantra that we can trust but must verify. Don’t blindly accept AI-generated outputs in the name of efficiency.

Whether or not the GenAI tools have delivered on the promise to create effective visualizations remains up for debate, but they can be used to generate code or a starting graphic with some guidance.

To get to a good chart, these tools often requiring a lot of prompting informed by a basic (or sometimes a more advanced) knowledge of data visualization design principles to provide direction on key design decisions like decluttering, swapping chart types, making purposeful use of color, and adding meaningful text.

Diver deeper into the use of GenAI tools in data viz design and interpretation in explorations from The Pudding (on creating a data story) and on Nightingale (on reading insights from a chart).

Yes, but our value proposition will change.

More people creating simple charts will reshape the value proposition for hiring dedicated data visualization designers.

Yes, there will continue to be a demand for well designed bar charts and line graphs, but the strategic value of data viz will be making the complex understandable. We may need to make a clearer case for the ROI for spending money on dedicated data visualization designers, which should nudge us to think deeply about how we communicate our value.

Organizations will likely make more of the basic charts in house as more people learn how to create basic charts, while we advise on and do more complex work that steps outside of that comfort zone. More of the subject matter experts building their own charts will put the creation of data visualization into the hands of those who are using those charts to inform decisions.

Established data viz designers can be thought partners for our clients, which is the structure many established design firms and consultants use in their work. We don’t take a design brief and work behind closed doors for weeks or months to build something, assuming we know best. Instead, we actively collaborate with our clients, prototyping ideas, sharing knowledge about why we’re recommending a particular approach, and iterating on ideas.

Shirley Wu’s story about her own mindset shift towards being a thought partner for a client (and sometimes accepting that what we love to create isn’t what our clients want) makes this case with a practical example that ends with a pie chart.

Is this shift good for our society?

Then, we can dig into whether more people creating data visualizations is good for our society as a whole. The potential downsides (particularly around amplifying misinformation) feel higher stakes, but I still believe they’re outweighed by the positives.

Yes, because charts are more widely available.

Charts show us patterns in data tables that would otherwise remain hidden. There is power in data visualization through that speed to insight.

Today, we see data visualization in our lives, not just at work and in the news. We see them every day scrolling through social media. We glance at fitness trackers and weather apps daily to understand our world and inform decisions like needing to take a walk or whether to pack an umbrella.

We also use visualizations to make sense of information on public health and climate change, where individual decisions impact communities. Having more accessible tools for creating charts has empowered a new wave of science communicators who can incorporate visualizations into their work without being data viz experts or learning to code. Often these charts are simple bars or lines. But creators also venture into graphics that help us connect to a topic.

In this fourth wave, we celebrate creativity and making complex information accessible visually. I often see isotype, frequency framing, and bubble charts that focus more on relative comparisons. By seeing novel chart types more frequently, we can build the mental models we need to read them and engage our audiences.

Yes, because we engage with data in novel spaces.

We see data in spaces beyond our devices. Data murals, data art, data sonification, data sculptures: how we engage with data has gone beyond visualization and shifted to enable entire data experiences, where you may not even realize you’re being immersed in a dataset. This is a good thing, as data can still feel intimidating and these creative forms can make data more inviting.

At the NASA Earth Information Center in Washington D.C., an immersive room helps you experience what it feels like to look down at the Earth from space through a narration and projection of stars and data. (Photo by Amanda Makulec, all rights reserved)

Showcasing data in community spaces through nontraditional projects opens up entirely new opportunities to bridge the data literacy gap, which remains a challenge as we tackle ways to help people be more careful consumers of data visualization.

Rahul Bhargava explores these spaces in his new book Community Data, featuring data sculptures and data theater implemented in even the most remote settings. Rather than brushing these forms off as ‘less serious’ or ‘less important’ than more traditional information graphics, these types of installations are starting to win awards like the Information is Beautiful Gold winner ‘Life Under Curfew.

Yes, but we need to be more careful readers of data visualizations.

Not everyone creating charts deeply understands the data they’re working with. Simple errors (or willful omissions) can result in more misinformation in the world, communicated in the language of charts. I’ve seen this first hand in my work in public health, particularly around COVID, wellness, and vaccine information. An attention-grabbing graphic quickly spun up in Canva that amplifies the findings of a poorly designed study hurts, rather than helps.

Designers can shape charts to communicate different messages depending on their audience and their goals. The same datasets used to make a case for mask wearing were reshaped and communicated across anti-mask communities during the peaks of covid. In “Viral Visualizations: How Coronavirus Skeptics Use Orthodox Data Practices to Promote Unorthodox Science Online” (MIT), researchers found that anti-mask communities often required that contributors create their own charts (distrusting those in the media). Knowing who created a chart influenced it’s trustworthiness.

As we face waves of mis- and disinformation influencing elections, policy positions, healthcare decision making, and more, we must be more critical readers of charts. We need to ensure everyone has the critical thinking skills to spot misleading graphics or those with a clear agenda.

Sometimes spotting errors is easy if you know where to looka bar chart axis that doesn’t start at zero, for example.

But other times, it’s being critical of what a chart is hiding and recognizing what isn’t there that is more importanta line chart showing only a limited period for the purpose of hiding the real shape of a curve, or summary statistics that fail to tell the underlying stories that live in the data when disaggregated by race/ethnicity or sex.

Alberto Cairo’s How Charts Lie is an excellent book on becoming a more critical reader of data visualizations for anyone curious about the subject (not just data experts).

With great power comes great responsibility.

As data viz creators, we need to be thoughtful about the what our charts show (or hide) and finding ways to make the complex more understandable. But as chart readers, we also have a responsibility to be curious.

Democracy only works for the greater good if we all take an active role in its systems of governance. The democratization of data visualization comes with great promise, but only if we work together as chart creators and readers, amplifying charts that inform and calling out those that mislead.

Next
Next

Finding a new career in data beyond international development