Sergey Nivens - Fotolia

4 features of great data visualization and storytelling

Data visualization and storytelling go hand in hand when it comes to explaining data. Here are four ways to make sure you build and tell a strong data story for your audience.

The great thing about data is that it can be used to prove anything, but how compelling is your data story? While technical issues like data quality and model accuracy matter, at some point you're going to have to explain what the data means.

If you're faced with a simple question such as "How many widgets did we sell last month?" a simple data visualization will do. However, when you need to explain a "why" question, you're probably going to have to create a data story.

"Data storytelling is the ability [to convey] the message to a large audience with a simple and clear message," said Charles Miglietti, CEO and co-founder of Toucan Toco, a data storytelling software company. "This is the very last mile of the data, and this is how you build true data literacy at scale."

Unlike a standalone data visualization, a data story has a beginning, a middle and an end. The beginning should explain what your report or presentation is about and why it matters. The middle explains those things in greater detail and the end summarizes the story and usually includes a recommendation.

Here are what experts say are the four most important features to successful data visualization and storytelling efforts.


All data storytelling requires context. This context has two elements: the point you're trying to make and the audience you're addressing.

Most likely, someone has asked you to explain why something has happened, such as why customers are churning. The question serves as the basis for your hypothesis and the data you'll use to answer the question.

"I like to start with what the problem was and why it's a problem," said Jeff Herman, lead data science instructor at the Flatiron School. "I found in the past a lot of people don't know what data I'm working with if it's a non-technical audience. [When I was] a data scientist for a railroad, we were looking at it as a month-to-month aggregate, so I'd show them the past month, then I'd go into my recommendations. There needs to be some sort of actionable take away from it."

The second important part of context is the audience, because the audience can influence the type of data visualization and storytelling language you use.

Aesthetics and types of visualizations

Data visualization tools provide a lot of freedom. However, creative license can work against clarity when clarity should be your goal.

The first thing you should think about [is] that data visualizations are not for yourself, they're for others.
Jeff HermanLead data science instructor, Flatiron School

"[T]he visualization that 'looks cool' is usually not the best visualization to use, especially when presenting to a non-technical audience," Herman said. "The first thing you should think about [is] that data visualizations are not for yourself, they're for others."

There are three mistakes people make when choosing a data visualization, each of which detracts from the clarity:

  • The visualization is too cluttered.
  • The color scheme is ineffective.
  • The visualization type doesn't fit the data or the audience.

If the visualization is too cluttered or too hard to read, people may misinterpret it, or you may lose their attention.

"Big companies like L'Oreal are very data-savvy for the most [part]," Miglietti said, "but we made them realize that going for simple visualization, good visual cues like color coding, a good legend and good context -- good message -- are much better, and it probably has better impact than providing complex visualization with a lot of features [or] a lot of information."

Colors should also be used in a way that advances clarity. For example, in a pie chart, are you trying to emphasize the greatest percentage, the smallest percentage or one of the midrange values? Use colors to enhance that connection, not detract from it.

Finally, one type of data visualization will make more sense to use than another. Some of the most popular data visualizations used for non-technical audiences include:

  • Pie charts, which show percentages;
  • Bar charts, which are good for comparisons;
  • Line charts, which show trends; and
  • Scatterplots, which show data distributions.

Data storytelling typically involves a series of data visualizations, which start at one point and end at another -- for example, from a general to specific or past to present.


Data stories include two types of narratives: data visualization narratives and the story narrative, which need to work together seamlessly.

The data visualization narrative explains what the chart or graph means.

"If is the first time [your audience is] seeing that data, everyone is going to need the narrative or it's going to take a long time to understand what the data means," said Bryan Coker, principal consultant of data and analytics at AIM Consulting Group.

The story narrative tells the entire story from beginning to end. It should have a logical flow that leads to a logical conclusion. A set of data visualizations alone doesn't tell an effective story because there's no transitional "glue" that holds them together. People naturally add those transitions when presenting. Remember to do the same when you're providing a written report or handout.


Insights have little value if they don't recommend action, such as whether to invest or divest. You've already explained what your data story is about and why it matters in some detail, but what should people do with this information? If cellular customers are churning because their calls are dropping too often, should the company build more cell towers or focus their marketing efforts only on certain areas?

The alternative is providing decision-makers with a set of choices and their associated probabilities so they can make an informed decision. Providing information is, after all, the goal of good data visualization and storytelling.

"It's thinking through the layers and making sure there's continuity between where you're starting and where you're ending," Coker said. "If you have the beginning and the end framed well, making sure getting from the beginning to end is done well is where big focus needs to happen."

Next Steps

Gartner predicts data storytelling will dominate BI by 2025

Dig Deeper on Business intelligence management

Data Management
Content Management