Whether you are a journalist or a marketer or an academic, you are going to need good data to tell persuasive stories.

Data can help to tell a story, illustrate trends over time, confirm anecdotal evidence, and lends accuracy to your narrative. But working with raw data can also be cumbersome and confusing. Unless you are a trained statistician, it can be difficult to look at spreadsheets and immediately recognize the discrepancies in the numbers or patterns that tell a story.

So, if you’re not a numbers person, how do you work with the data and numbers to discover whether there is a story inside all those figures and variables? To get some clarity, and some peace of mind, here are a few steps to get you going in the right direction.

1. Know Your Audience

When writing, you should know who your audience is by trying to anticipate who will be reading your copy. Likewise, you need to bear your audience in mind when you’re deciding whether or not to seek out which data to analyze, and whether you should do the analysis itself. The reason is that you want to ensure the data is relevant to your audience. Are they experts in the field your story represents or are they novices? Will the data reveal something that will help them or solve a problem, or will it just tell them something they already know?

Good data storytelling needs to provide value to the audience so, before you start, figure out what that value might be. Doing this will help you decide which data you use for your research.

2. Get the Right Data

Your story is only as good as your numbers. Use your critical thinking skills and make sure that your data comes from reliable sources. Trying to crunch police data? Pull them from the police department directly, but also look around for additional sources, like the FBI or your local university criminology department; both will likely have numbers that are even more granular than that which the local police department releases.

The same is true for other forms of data. You want to make sure the data you are using directly reflects the questions you want to be answered. You won’t be helped by data that is more general. Instead, you want to find accurate and niche data from reliable, non-partisan sources (government offices, universities, non-profit foundations) that directly address the story you are exploring.

If you are using consumer survey data or similar, look at the original set of questions. Are they written in a way to skew the answers in a particular direction or are they written to solicit answers that are more divided? If you don’t know, work with your survey provider to make sure you are asking the questions right.

3. Make Sure Your Data Set is Clean and Organized

Delete rows of missing data. If the data jumps strangely high in either direction, make sure it’s not an error. If you are working with a survey question, make sure the question is in a column and all the answer choices are clearly stated across the rows. Are there choices that you don’t care about and aren’t necessary for your story? If so, you can delete them to ensure that you do not use them in error.

---

---

4. Examine the Relationships Between Data Points

Highlight the data that you know you are interested in and just focus on that. Scope out patterns that start to emerge. Does the data correlate, meaning that are there two or more variables that could show a positive or negative relationship? Do you see an ascending or descending order, or any other pattern, or does the data basically stay the same? By focusing specifically on the set of data you need for your story you’ll have a better chance to spot those trends.

Finally, check for outliers, which are data that are or behave differently to every other number in the spreadsheet. The numbers should have some common sense logic to them, and even if they don’t, there should be a clear-cut reason to explain why some numbers are clearly out of the ordinary. If you can’t find those reasons, make sure the numbers aren’t an aberration or don’t use them.

5. Be Willing to Walk Away

Sometimes the numbers don’t say anything. If your survey data shows that everyone loves ice cream, no matter their age, household income, gender, the region of the country, and so on, there is nothing to write about. Be willing to walk away from a dataset that shows no divergence between numbers. Don’t censor or manipulate the data, either. If the data tells you something you already know, walk away and find a more specific angle to the story you’re interested in learning more about.

---

So next time you are asked to put a story together based on data, do some audience analysis to help you find the appropriate data to gather. After doing so, ensure that you are using niche enough data and that it’s not coming from biased sources. When you have the correct and appropriate data, you can work towards ensuring you keep it organized and clean. Lastly, check for outliers in your results and the relationships between data points as this will help you zero in on a story.

Again, if your results don’t tell your team a story that you feel would be worth bringing into fruition, even though it may have eaten up a fair bit of your budget, it’s better to start over and go at it again from a different angle. You will be much better off having an interesting story that has correct and valuable data behind it compared to one that’s been skewed to make it look more exciting.

---