Data-Driven Storytelling Builds Reader Trust

These are hard times for people who prescribe to expertise, data and facts. Social media – especially Facebook and X/Twitter – has mutated from a cute place to share puppy photos into a seething cauldron of name-calling, Do Your Own Research and A Collection of Dumb Videos and Hearsay.

In response to that rabble, I offer this T-shirt slogan:

As a data-driven content Dude named Deming once said, "Take the data, leave the cannoli."

If you don't read T-shirts, it says, “Without data, you’re just another person with an opinion.” (I wasn’t the first to say it — that was W. Edwards Deming.)

Or I just quote The Big Lebowski (shown above). Of course, T-shirts don't win arguments. Neither does bluster. You've got to have the facts (and data) on your side.

Data-Driven Content Cuts Through the Media Echo Chamber

For more than 20 years, I’ve used data storytelling to stand out from the crowing hacks with opinions. In fact, finding and using data as a data journalist has been my secret superpower. After leaving traditional journalism, I found that data analysis can be foundational in media relations, content creation, blogging and writing.

Data, if used properly, provides the perspective readers need to be more completely informed. Verified, accurate data can also help build trust with readers, who, justifiably, have become more than a bit leery of outrageous claims lately.

What’s more, since newsrooms are laying off data journalists, that presents an opportunity for you and your organization. You, too, can be a data journalist!* (* = It would help to be a regular journalist first, though.)

I’ve used spreadsheets, database managers and mapping software to find:

—     a DUI hot spot around the Dallas Cowboys’ and the Texas Rangers stadiums, which is one of the biggest party spots in Texas, booze-wise (the analysis was for a DUI plaintiff lawyer).

—     that contrary to urban legend, metro Providence, Rhode Island, has more restaurants per capita than any other metropolitan area in the United States (published on Yahoo! News).

—     dozens of troubled trucking firms for my legal clients, who represent trucking firms with regulatory compliance.

From the outset, the notion of using data to tell stories can be off-putting. What if I get it wrong? Isn’t using data just a more sophisticated way to lie? Where do I even start?

The reluctance to get into data analysis is actually healthy. A lot can go wrong, so that instinctive hesitation is good. But seeking out and using data shouldn't paralyze one with fear.

Start with a statistic or number that catches your attention and go from there. Often, when you find one number, someone’s tracking things. If there are enough numbers to tie together, you’ve got a chart. If you've got fixed geographic points, you might be able to create a map, or even a heatmap.

Regardless, whatever data visualization that results needs to tell a compelling story. You're looking for a hockey-stick spike. Or maybe a scatterplot with some interesting outliers. Anything that makes a reader want to understand more.

Here are two statistical hockey sticks (apologies to Don Rickles):

I found these stats above while researching homeowners’ associations for my client HomeEc, which has been grappling with HOAs banning the construction of tiny homes within their subdivision: In short, that chart presents a compelling argument for codifying the construction of tiny homes (also known as accessory dwelling units, or ADUs) across the nation.

Using Common Sense to Find a Statistical Hockey Stick

I generated the next chart while working with the Safe Healthy Playing Fields organization, which promotes the use of natural turf, rather than synthetic materials that must eventually be dumped when it wears out.

Artificial turf, I've observed, was barely around when I was growing up in the 1970s. It was easy to deduce that the number of these fields have shot through the roof, based on the prevalence of the fields now.

In the case of artificial turf fields, Excel even projected out how the number of artificial surfaces would proliferate in the U.S., from one in 1966, up to 20,000 by the year 2031-ish.

You Don't Need a Ph.D. to Create Data-Driven Content

Both charts speak more loudly than a stadium full of blowhards at maximum Personal Opinion.

Neither demanded advanced knowledge of data or mathematics. They just took a little research, with a few ham-fisted Google queries. The HOA data actually came from an HOA advocacy group. So, yes, Virginia, you can use your opponents’ own data against them.

In other cases, there are entire pages of datasets to download, which I did for my story on which metro areas have the most restaurants per capita. That analysis was more complicated – I joined bureau of labor restaurant stats with Census data to generate that table. But maybe I’ll tackle that in a later blog installment.

5 Things to Remember Before Unleashing Your Data-Driven Content Tsunami

While writing about numerically driven trends doesn't require rocket science, it does require some caveats:

• Always cite your source, and include it with your data for review;

• When searching for data, keep your eyes out for other data sets that might come in handy for future analysis;

• Create a written description (called a "nerd box") of how you generated your chart, so you and others can retrace your path;

• Make sure others can reproduce your results; and,

• If you’re new to using data, work with someone who is experienced in it, to prevent mistakes.

Bulletproof Your Data So You Can Stand Behind It

Mark Twain memorialized the phrase: “There are three kinds of lies: lies, damned lies, and statistics.”

And yes, there are practitioners of bad statistics/data. My favorite example is newspaper editors who publish charts that don’t start at “0.” Those charts show a sudden huge leap in a particular area, without clearly spelling out to the reader that they’re looking at only a small sample of data. More recently, I saw a data journalist from an outlet called "Insider Monkey" identify New York City as having the most restaurants per capita. That surprised me, because in my analysis, NYC finished seventh. When I checked her numbers, it appears she disregarded higher numbers from San Francisco to make NYC No. 1. I've contacted her with my observations. Here's what I got back:

When posting data, make sure you're ready for "feedback" from gadflies by bullet-proofing your data.

Those are the risks you take when sticking your neck out to produce an eye-bulging chart. Your numbers MUST be bulletproof and irrefutable. Even if you work at someplace called "Insider Monkey." (Of course, I've written for "Yahoo!," so who am I to judge?)

Quality Data Analyses Can Produce More Than a Good-Looking Hockey Stick

In my capacity as a data journalist, I once showed a police chief statistics that revealed his officers were arresting a vastly disproportionate number of Black children in the community. His answer? “You can make numbers do whatever you want.”

That kind of data illiteracy, or that claim of data illiteracy, is truly disheartening. Those numbers came from his own department (I just analyzed them). Yet he didn’t believe them. He didn’t see the problem the data represented.

Regardless, you must respect readers’ intelligence. Let your numbers tell their own stories -- this is the heart of data storytelling. If the numbers don't do what you want, move on, without any chicanery. In doing so, you can become a trusted source, when those are becoming increasingly scarce nowadays.

There’s enough online chicanery already, isn’t there?

Data Storytelling Plans