Bar charts and line charts are the most popular ways to visualize your data. And for a good reason. Thanks to tools like Excel and Google Sheets, they are quick for anyone to make and familiar and understandable to most audiences. Want to show growth in your user base? Love that up-and-to-the-right arrow? These visualizations work great.
But sometimes – or dare I say, most of the time – your data can tell you more when visualized in different ways. While those easy-to-use tools lower the barrier to entry, they don’t serve up enough different types of charts or options for those charts. As a result, when we talk about visualizations like connected scatterplots or sankey diagrams, most people are intimidated because they’re unfamiliar with them. And when it comes to those reliable bar charts and scatterplots, we aren’t familiar enough with their advanced options and how to design them to make truly impactful data visualization.
I’ve been making data visualization for a long time and have been thinking about how to make it easier for folks to get comfortable with more ways to visualize data. Here are 6 tips to help you get started so you can expand the possibility of what you measure.
1. Know Your Goal
Charts are a form of communication, which means you need to understand what you’re trying to communicate and who you’re trying to communicate to, and why it’s important. Are you trying to Explore the data or Explain the data? Are you trying to discover patterns in the data? Then you’re the audience. Are you trying to show patterns in the data to others? Then you need to be able to describe that audience and their goals.
If you’re interested in how different roles in data organizations use data visualization differently, you should check out my video series on Designing for the Data Visualization Lifecycle, where I discuss in-depth how data engineers, data scientists, data analysts and downstream data consumers approach data visualization with different goals and assumptions.
2. Explore With Multiple Points of View
Exploratory Data Analysis (EDA) should leverage views that show trends and correlation. That means using line charts and scatterplots a lot, but also more advanced charts like funnel diagrams and network diagrams.
But it’s not just about using a kind of chart, it’s providing yourself with context. If you’re using a product like Noteable’s Data Explorer (DEX), lean into the functionality it provides natively to see small multiples of the data using composite charts like a Scatterplot Matrix. This gives you an excellent overview of your dataset to see how different metrics correlate.
But you don’t have to use small multiples like this to look at correlation as an overview. In data science, it’s common to facet the dataset by metric or dimension to show the same chart across categories or metrics. DEX exposes this faceting for almost every chart, making it easy to pivot on a metric or dimension to quickly scan your charts to find interesting patterns.
Finally, when exploring you should also push yourself to find non-numeric patterns using more complex charting methods like network diagrams and funnel charts. Insights don’t only exist in the numerical attributes of your data and if you’re not looking for geographic, flow or network patterns, then you might miss the key factor in understanding your data.
3. Explain with Engagement
When delivering charts for explanatory data graphics, you need to understand your audience. Not just their goals but what they’re comfortable with and how they typically present their data. When exploring data, you might be comfortable with chart types or chart options that will intimidate your audiences. The best way to deliver effective charts is to deliver them in a way your audience is comfortable with. That doesn’t mean they need to be boring, though. Here’s a boring and effective information display.
Bar charts are great for numerical precision and there’s nothing technically wrong with this chart. But take that bar chart and use just a touch of design to draw attention to significant trends in a way that the default presentation may not provide.
See how simple color rules, bar labels, and annotation have changed a basic bar chart into a story about countries with more or less freedom than the United States. I put this together in a couple of minutes with no intent other than to demonstrate the effectiveness of color and labeling, but even that tiny amount of effort made a chart that is instantly more engaging and impactful. Just think about what you can do with a little more time and a real story.
It’s not enough to use the right display for the audience and decorate it to make it more engaging and insightful. Just like with exploration, context is key. Use multiple views into the data to provide your audience with that context. That’s why DEX allows you to build dashboards out of your datasets to show multiple views into your data simultaneously.
The great thing about context is that it pushes your audience out of their comfort zone. You can provide them with a few views that are familiar (like bar charts and scatterplots) and then hit them with a more exotic chart like a treemap which provides them with a new layer of insight and simultaneously increases their familiarity and comfort with the new chart types.
4. Decoration Isn’t Just for the Home
We’ve seen how color can help make a simple bar chart into a story. Good tools like DEX expose advanced functionality that allows you to take a simple chart like a scatterplot and make it more accessible and information-dense. Compare these two scatterplots, one that just plots two metrics and the other that uses more advanced functionality to make a rich information display.
We’ve used color already but there’s more going on here that you might not even notice.
Title, Subtitle and Footer – Rather than make the reader guess what’s going on, you should situate them within the data display by stating the proposition explored with the data and give them something that they always need: context. Use titles to explicitly state the claim of the chart, subtitles to prompt them on what to think about and footers to provide more detail like data provenance.
Advanced Encodings – Good tools like DEX expose additional options such as trendlines, marginal graphics and circle size encodings (to make a graduated symbol plot) to not only encode more data but to make the data presented more approachable. Now you can quickly see the distribution of the data along each axis and understand the trend numerically so you don’t offload that pattern onto your user, freeing them up to think about more complex patterns.
Conditional Formatting – By using color to encode another metric (Corruption) we can present possible explanations as to why the two metrics don’t perfectly correlate but also prompt discussions of what other metrics might explain what’s going on.
Annotations – Decorating an outlier or important datapoint with a natural language callout like our annotation of Myanmar makes the display more engaging and gives more–you guessed it–context to the data.
5. Everyone Loves Maps
Why does the Terrible Maps twitter account have 1.7 million followers and the Terrible Networks twitter account not even exist? Because people know and understand maps so well that we can all make fun of a bad map together, whereas network diagrams are still hard enough to make that we can’t even agree on what makes them suitable.
So whether you’re using point maps or choropleth maps, if your data has a geographic characteristic, go the extra mile to present that data geographically. Maps are much more complex than bar charts but we don’t realize it because most of us learn to read maps at a very early age. That means they’re familiar and engaging and it empowers your readers to find patterns that may exist in the data in a way that’s obscured in a scatterplot or line chart.
Remember that maps are also subject to the Explore/Explain approach to making data visualization. Some maps, like this point map, might be better suited to exploring the data than explaining it, so just because you’re using a map doesn’t mean every map is perfect for every use case.
6. Just Because You Can Doesn’t Mean You Should
The final tip isn’t necessary with more limited applications, because you just can’t make rich charts with underpowered tools. But remember that even though tools like DEX provide all this functionality, you don’t need to turn everything on at once.
Remember, especially with explaining data, encoding more data into a display doesn’t mean you’re enabling your users more. By turning on every feature all at once, you might turn your chart into something that actively prevents insight. Data visualization is about communication and so let me leave you with a tip from one of the great communicators in the English language:
“I didn’t have time to write you a short letter, so I wrote you a long one.” – Mark Twain
Easily Visualize Your Data with DEX
I used Noteable’s Data Explorer (“DEX”) to make all the charts you saw in this article. DEX is a powerful tool exposing many data visualization modes and features you don’t see in traditional no-code applications like BI tools. DEX gives you options like annotations, complex charts and extensive functionality within the chart to use multi-axis displays or faceting like you’ve seen above. That gives you great power, but remember the immortal words of Uncle Ben!
Create a free Noteable account and start using DEX today.