The Importance of Storytelling and how to effectively use it in Dashboards

Working as a data analyst in an MNC, I was tasked to build a dashboard to show the performance of a machine learning algorithm. I spent the next few days identifying the data points and tables I’d require and curated my SQL query for data extraction. Everything was ready, and I built the dashboard by showing the summary of the algorithm’s performance, the variables that had the heaviest weightage to the numbers, and other visualizations that would show its general performance.

With the dashboard built, I reached out to my manager for a quick review of the dashboard. He gave a quick glance at it and asked one question which changed my perspective of dashboards. That very question was…

“if I need to tell a story that will lead to changes made to the algorithm, how would you do it with this dashboard?”

That is when I realized, that even though the visualizations highlight key insights to the algorithm, the insights were all over the dashboard, making it extremely difficult to piece everything together.

Storytelling with data

After the conversation with my manager, I reflected on how I can do better with future dashboards. This led me to discover a few things that I can improve my dashboard-building skills.

Constantly asking questions on the insight/data

“Root cause analysis” is the act of constantly asking why until you understand the root cause of your problems (if any). A key figure gives you a good summary of the situation, but the act of asking why usually leads you to the goldmine. I use the 5W 1H format to guide me to ask more effective questions. Let us run through an example:

Assuming that the key figure is showing that it is 9% less accurate than last month. My initial thoughts could be:

1. What caused the key figure to decrease this month?
   
2. What are the main contributors to this current month’s forecast?

As I get answers to my initial questions, I will ask another set of questions relating to the earlier questions. Some examples could be like:

1. Oh. It’s because of x, that caused the key figure to be 9% less than the previous month. Has x been consistently pulling the numbers down in recent months?
   
2. Does x have a significant weightage in the algorithm numbers compared to the other variables?
   
3. What is happening with x, that is causing it to have a negative effect on the key figure?
   

The action of constantly asking why will give you a better understanding of why the key insight is in a certain way. However, it will usually lead to having too much information. This leads me to my second point, piecing them up together.

Piecing the answers together

With all the answers that you’ve gathered, it is time to piece them together. Something I do that helps me is putting everything into a notepad, and then start joining them together to create a story. One example would be:

1. I would like to know the overall summary of the key figures.
   
2. Once I know the key figure, the next thing I want to know is what is causing the overall summary to be such
   
3. I then want to know if this has been consistent over the past 3 months

Because there will be many possible stories that have a good flow, I like to write all of them down. I’ll then proceed to review the “stories” once I have a list. The story that is the best fit for the dashboard’s agenda will then be chosen as the “chosen one”. With the chosen storyline, we can move on to the final step. The creation of the story on your dashboard.

Curating the story

Let me ask you guys this question. Do you think it is easier to

1. read a book where the words are left to right, and then move to the next space whenever required
   
2. Read a book where the words are all over the place, and you have to piece them together.

I would think the majority of you prefer option #1. Similarly, your dashboards should be structured like a storybook so that the flow is consistent as you lead them to the main points. This means having the main question answered at the top left of the dashboard, with the supporting questions answered following.

Summary

Creating a dashboard is easy, but creating one that leads your users to insights far more than just a number takes practice. Using the 3 steps that I’ve highlighted above, creating a dashboard that tells a story becomes easier over time. Understanding this has made my dashboards more effective for my stakeholders, which usually leads to praise by them. Try it for yourself, and let me know if you find this effective for you.

---

Thanks for taking some time to read my post. I am a data analyst in an MNC that is dealing with supply chain-related data. I’ve curated this internet persona to share my journey as a data analyst, and my learnings to the public so that everyone can learn and improve from my discoveries and mistakes. If you do enjoy this kind of content, be sure to leave a comment down below to let me know your thoughts, or reach out to me via Instagram to have a chat regarding anything.