The newly evolving field of data science can be made more engaging, informative, and creative using visual data storytelling which makes insights more memorable. Insights are the non-obvious valuable facts, and this blog is a walk-through on what is visual data storytelling, how to create data stories, and what are the contents of a data story.
Case studies, Toolkits, Frameworks, and Methodologies to Drive Data Storytelling At An Org Level
Download Whitepaper on Data Storytelling
Table of Contents
Let us tell you an interesting anecdote to understand why we need storytelling.
Rob Walker and Joshua Glenn auctioned a bunch of items that they bought from eBay on significantobjects.com. They spent about $250 and asked people to write stories about each of these objects.
For example, a person described it as a small statue that stood on his front hall window cell of his favorite aunt’s home. And the story goes on to describe the history behind the object. They bought it for $2, and they sold it for $50, which was a giant leap!
All the products were purchased for $250, while they got sold for $8,000. This anecdote demonstrates the power of storytelling in terms of revenue gains – a whopping 3000%!
There are two main reasons why stories sell:
Therefore, engaging the audience based on their interests and sharing information as a story has immense power.
The starting point is you have data with you, but what do you do with it?
The most important thing is to know who your audience is because they determine the structure of your data story and analyses.
You can work on the same dataset, e.g., say sales data. Let’s suppose there’s a need to predict the subsequent quarter sales. The head of products will ask which products grew the most and the sales head is probably interested to know if the targets were achieved or not. Each group has its own set of questions, and they are not as much interested in other’s questions.
So, without knowing who your audience is, you would not crack the problem that interests them. If you are starting with the analysis right away, you are your audience, and if it’s someone else, find out who it is and write it down.
There is a specific way in which you can write it down:
1. Mention their role
Role:_______________ (Be specific – “Head of Sales,” not “Executive.”)
2. Mention their name
Name a natural person – “Jim Fry,” not “Sales Head.”
To summarise, start by knowing who your audience is. And then move on to understand the problem because the audience’s problem defines your analysis, which determines your story. Effectively for each person, you want to answer four questions:
There is a reason why each of these is important; the situation will help you understand what has happened. If we know the impact, firstly, it is easier to prioritize, and secondly, we can figure out whether we can achieve the effect.
While writing, make sure you write two sentences before you start with any data analysis and data storytelling. Also, write it in a structure –
“(Person, Role) is in (situation), and faces this (problem). By taking (action), she can drive (impact).”
Now let’s take some real-life examples like, say, commodities.
Adam heads the purchasing team of a leading European brewery. Their plants had purchased items from several vendors, and discounts were low. The number of weekly orders was high, increasing logistics costs. But he didn’t know which plants and commodities were creating issues.
When he reached out, every plant denied it. The action he could take was to consolidate all of the vendors and reduce order frequency. Then they could increase their discounts and reduce logistics costs. That was his impact. By knowing this, we have a more precise definition of the problem.
Chris, who heads the operations team for a US airline, encountered a similar issue. He has an SLA to deliver cargo from the flight to the warehouse in less than 1.5 hours. Unfortunately, that is 15% lower than their current best performance.
Chris didn’t know why they were getting delayed? Is it because people are a problem, the assets are a problem, or the type of cargo is a problem? Focusing only on the right area of concern, Chris can reduce the turnaround time reducing the spending.
The next step in visual data storytelling is actually to perform the correct analysis. Now, this is a somewhat complex and messy decision tree. But there is a simpler version of a decision tree that helps you understand different ways to discover the kind of analysis to solve a given problem.
And the reality is – There are many ways for people who are familiar with data. You might already know some processes to perform analysis, but first, study the circumstances and go for decision trees, structures, or a flowchart, one that fits best.
What is the type of outcome? Is it continuous or categorical?
Continuous involves numbers, whereas categorical has labels like high, medium, low or red, blue, green. Suppose we move ahead with Continuous outcome.
How many predictor variables do you have? Well, Chris may tend to have several variables from the airline’s perspective. Do I have enough forklifts, have enough trained staff, have enough team members, and have enough inventories for the cargo to be sent? That is two or more variables.
What are the types of predictors? Well, some of them are continuous types and some categorical.
In that case, Chris could use multiple regressions to find out which is the better predictor of an outcome or use analysis of covariance (ANCOVA). Once you know the nature of the problem, it becomes easier to figure out possible ways. We recommend that you don’t go listing all of your possible types of analysis. Use a decision tree kind of an approach, and if you want to extend it with other methods, you can search and fit them into the slots. Finally, you have the result, and the next step is converting that analysis into an insight.
An Insight has three criteria –
👉 First, it must be big – we want to change the outcome substantially.
👉 Second, remember in one of the earlier examples, we had told the farmer that the farmer should increase rainfall. Well, the farmer cannot act on it. An analysis is useful only when people can act upon it. Then, you can call it insight.
👉 Thirdly, it would be surprising if you told a farmer that rainfall will increase the yield or if you told the head of finance that to increase your profitability, you need to cut costs or increase revenues. They might end up laughing at this obvious analysis.
The analysis must be big, useful, and surprising. If it meets all three criteria, then it’s an insight. If it doesn’t, it’s not.
The lesson we learn from the above examples is to find the correct analysis based on the problem.
Step 1: Use a flowchart and figure out the top analysis you should focus on.
Step 2: Make sure you filter it based on three criteria: is it big, is it useful, is it surprising? At this point, you have an insight.
Step 3: Next, begin building a story.
The most important thing is to start with a takeaway, one sentence that summarizes your entire story. Start with the moral of the story as the key takeaway – this is effectively the elevator pitch around which everything revolves. Here’s a simple checklist that you can use to figure out if you chose the takeaway right –
Construct a pyramid or tree-like outline –
Supporting points build the credibility of the statement. The statement becomes less credible if there are no supporting points. These help you build a storyline. It’s not the whole story yet, but it’s a storyline.
Now let’s learn how to present a story. You should make two choices; the first is what medium you want to show? And what is the design of that medium?
Different storytelling formats allow you to create, narrate, and enable people to explore alternative narratives such as interactive dashboards to blend more than one design. And what kind of chart to choose?
That’s again a function of what you want, and there are three things that you can consider –
You want to help people explore the data or showcase performance; you want to tell people that one option is doing well and the other is not. Or you may want to go to the next level and tell people why it is doing well or not doing well. One way of thinking about it is, let people figure it out, or you tell them what’s wrong, or you go further and tell them why it’s wrong.
The vast majority of results that we want to produce fall into one of the above three choices. Now how do you do that? You want to help people explore the data, then again help them show the magnitude where the chart flicker would help.
After doing all this, we finally have the audience with a set of problems whose data we have analyzed and converted into insight. Further, convert it into a storyline that is presented as a medium using a specific visual design.
It is also essential to annotate these. Try out one of the 72 types of data visualizations for data storytelling we compiled from our repository.
When you add these four types of annotations to any chart, it becomes incredibly readable, far more informative to the audience, and they will make decisions based on these.
With this, what you have is the structural ability of the framework to go from data to a dataset and finally create a data story.
Remember, data stories are memorable, viral and they have a 30X return on investment! Keep enjoying your data stories and stay connected to explore in-depth.
Visual Data storytelling is missing from business communication. Number crunching job roles such as data scientists and analysts can seldom make the stakeholders understand the numbers. Stories are memorable, viral, and send an actionable message.
With this blog, we would like to spark the need for visual data storytelling among the members of the data community. We offer hands-on experiences to data scientists and analysts and teach to create insightful data stories.
Check out the agenda, process, teachers, and benefits of our data storytelling workshop.
If you like this article, do share it with your fellow data enthusiasts. Finally, make a data story and tag us on social media (@gramener).
In today’s fast-paced world of e-commerce and supply chain logistics, warehouses are more than just… Read More
What does it mean to redefine the future of manufacturing with AI? At the heart… Read More
In 2022, Americans spent USD 4.5 trillion on healthcare or USD 13,493 per person, a… Read More
In the rush to adopt generative AI, companies are encountering an unforeseen obstacle: skyrocketing computing… Read More
AI in Manufacturing: Drastically Boosting Quality Control Imagine the factory floors are active with precision… Read More
Did you know the smart factory market is expected to grow significantly over the next… Read More
This website uses cookies.
Leave a Comment