Storytelling with Data – Part One

7 min read
Storytelling with data part one

Storytelling as a Data Scientist

Storytelling, in itself, is a fundamental skill for a data scientist to have in their tool arsenal. Storytelling, with the addition of data, adds another layer of credibility that will reinforce your narrative. To provide an analogy for this concept, we can refer to the data as the icing on our metaphorical cake. One could also think of it as the filling to the pie. However, when data is involved, the cardinal rule that highlights the novices from the experts is the knowledge that pie charts are useless in the world of data science. Pies should solely be vesicles for an overabundance of overripe fruit and not by any means a way to break down percentage comparisons, but more on that later.

It can be argued that data scientists are drawn to their profession because of their passion in uncovering the stories behind the data. This, of course, is an oversimplified explanation of what a data scientist does, but as Dr. Brené Brown famously explained during her TEDx presentation that “Stories are just data with a soul.” Poignantly in this simple sentence, Dr. Brené Brown exposed the reason behind why data scientists are drawn to the data. Unstructured data is ubiquitous in our daily lives, and we endlessly and subconsciously sift through it to make better decisions for ourselves. If you can digest for a moment what Dr. Brown is trying to convey, you can appreciate her in-depth understanding of how stories have the ability to inspire a call to action from their audience.

Why is the Story so Important?

important storytellingStorytelling experts Joe Lazauskas and Shane Snow explain how stories affect us at the molecular level. Stressful stories can increase pulse rate and convince the listener to secrete cortisol, the stress hormone. Other stories can draw the listener in and cause the synthesizing of oxytocin; the empathy hormone is also referred to as the “love hormone.” When our brains release oxytocin, this signals that the listener has formed a psychological connection with the story and is emotionally invested in the outcome. This latter scenario is ideal when trying to convince your audience to believe in and to feel motivated by what you are pitching.

Know the Difference Between Appropriate and Awful Visualizations

Visualization DifferencesVisualizations in themselves are vital to storytelling when appropriately used. We have all been in the situation where an acquaintance shows you what seems like a never-ending pile of pictures from their exotic vacation. The first picture might be vaguely intriguing, but by picture number three, it becomes increasingly difficult to feign interest. By the time the last picture rolls around, you have no idea what you are looking at and have been mentally checked out for a while. Would you be able to repeat what you have just seen and heard? Certainly not. The same goes for storytelling; one cannot just present a meaningless visual after another and hope the narrative lands. The only thing that will be landing firmly there is your unfortunate failure in conveying your story. Storytelling with data represented by visualizations should provide meaningful context and further validate the importance of your story. The data should be a tool for connecting the audience to the story and to communicate on a deeper level why they should care. This is why one meaningless visualization after another is consequently so ineffective and can have severe repercussions when presented improperly. Successful storytelling is really just the ability to inspire some sort of reaction.  Generally, positive reactions such as a change of heart, an emotional connection, or call to action are preferred. Adding data to the story enriches it by furthering the narrative and providing concrete goals of showing the current location of “x” and the goal of reaching “y.”

Know Your Audience

appropriate data for audiencePersuasive storytelling with data begins with knowing your audience and understanding their technical level. Are you presenting your findings to a group of fellow data scientists in your organization who will appreciate your industry specific lingo and will be enthralled with your mathematically heavy analysis of your findings? Or perhaps you are presenting to your Homeowner’s Association, which includes a diverse group of people with varying backgrounds on why you should change landscaping companies to keep your monthly HOA fees from increasing. In this latter scenario, it is pertinent to tailor your storytelling and data visualizations so that every single person present will understand the point you are trying to convey. There will be no need for anything other than simple visualizations that can be easily crafted in Excel. You will find this to be more than efficient even if you are capable of producing the most visually stunning dashboards on Tableau or Power BI.

Components of Storytelling with Data

data storytelling componentsCole Nussbaumer Knaflic is arguably the leading expert on storytelling with data and wrote a book on this topic aptly called Storytelling with Data. As a former Google employee, Nussbaumer Knaflic was tasked with leading workshops on how to effectively communicate with data within the organization. Her workshops were so successful that they led her to a publishing deal and provided the opportunity to teach her workshops globally to a variety of organizations. Nussbaumer Knaflic breaks down the components of compelling storytelling with data into six guidelines:

  1. Understand the importance of the context and the narrative you wish to convey.

 It cannot be emphasized enough how important it is to know whom you are communicating your story to. You have to put yourself in the shoes of your audience and figure out what makes them tick. You have to decide which points you must use to harness their empathetic nature and choose the appropriate data and visuals to support these points.

  1. Create visuals that are effective and appropriate.

 Choosing effective visuals really just boils down to allowing yourself to be a snob. You can, and you must be very selective about your choice. Nussbaumer Knaflic recommends line graphs, bar graphs, tables, areas, and heatmaps just to name a few. She does not, however, recommend pie charts in any shape or form (here is looking at you, a donut chart.)  Her argument against pie charts stems from the fact that the brain has trouble interpreting them as it is hard to compare the angles without a scale. Please do not feel obligated to limit yourself to solely static visuals. Hans Rosling’s 200 Countries, 200 Years, 4 Minutes perfectly illustrates how effective adding movement to the data allows the audience to visually sift through and perceive the data.

  1. Declutter.

It is highly important to not fall into the trap of trying to cram too much information into a single visual. Nussbaumer Knaflic explains the human brain has a finite amount of processing power, and when given the choice of sifting through a needlessly complicated visual or mentally checking out, the brain will choose the latter. Effective storytelling with data requires the audience to be engaged in the story throughout the beginning, middle, and end. It is highly detrimental to the effectiveness of your narrative if your audience is not engaged in its entirety.

  1. Home in the audience’s focus to exactly where you would like the attention.

 After you have decluttered, the next item is to make sure you have the attention exactly where you would like it. You can do this by leveraging the brain’s subconscious iconic memory by using preattentive attributes. This is also the time to make sure your visuals appeal to not only the eyes but also to the empathetic side of the audience. In other words, it is time to ensure the oxytocin is flowing.

  1. Implement design that is accessible.

 In order to implement an accessible design, Nussbaumer Knaflic argues that you have to think like a designer. This circles back to the concept of knowing your audience and understanding how they think. In knowing this, you will be able to predict how they will interact and respond to your created visuals.

  1. Ensure you are telling the story you wish to convey.

This step seems to be the most obvious yet realistically the most crucial component of storytelling with data. Think back on stories that have resonated with you the most. It would not be surprising if some of these stories are those you came into contact with years ago, yet you can still remember every component in great detail. Engage your audience through being your authentic self. Know that the story you are telling is not for you, but for whomever, you are telling it to.

Storytelling with Data: USD’s MS-ADS Program

USD data science storytellingThe ability to be a successful and succinct storyteller with data will be inexplicably one of the most valuable tools you will have as a data scientist. As a data scientist, your company will count on you to tell data-driven stories that can be readily understood and inspire a call to action by key stakeholders. It will be your job to shine a light on the problem and inspire a response. Data visualization guru Stephen Few explained the concept of storytelling with data by saying, “Numbers have an important story to tell. They rely on you to give them a clear and convincing voice.” In Storytelling with Data Part 2 we will learn how to harness the audience’s attention and create data-centric visualizations.

The University of San Diego Master of Science in Applied Data Science (MS-ADS) program offers a robust curriculum that focuses both on the technical and soft skills required to become a successful data scientist. USD’s MS-ADS online program ensures that our students graduate with the data scientist skills needed to succeed in the workforce, and understand this skillset is not limited to strictly technical training. USD’s MS-ADS online program was developed by industry experts who have firsthand knowledge of the practical skills required, and have designed its courses to reflect this ever-evolving industry.

Considering Earning Your Master’s in Data Science?

Free checklist helps you compare programs, select one that’s ideal for you.