The essential data science skill everyone needs

Image for post

What is a story?

A story is something that has three core connected elements. A starting state setting up or context, then some challenge or conflict in the middle, and finally, a changed state, resolution at the end.

We can call them the 3 C’s:

  • Context
  • Challenge
  • Conclusion

Stories can help with people’s comprehension and recall. They can provide important context and in certain cases, even foster empathy and emotional connection to the findings. In essence, stories activate and engage more of our brain’s capacity than simply looking at the numbers alone.

Of course, like any other powerful tool, there are potential pros and cons to using stories. The cons include distorting the data with meaningless patterns and false positives to make it seems like good data.

Therefore, to avoid these mistakes you have to make your stories clear, concise, and accurate. It’s also crucial to be selective and eliminate details that are meaningless or misleading. You can read more about how to make effective visualizations in my article 5 Best Practices for Data Visualization.


Creating your Data Storytelling

Here, we will walk through all steps to create a story using the Superstore Sales Dataset. The data contains 21 columns and a few lookup tables for returns and managers. Besides, additional columns were added to complete the dataset. The whole process was made with Tableau but you can use your favorite tool to create your stories.

The process will be divided into 4 steps:

1: Create a design checklist to guide the exploration of the data and the design decisions for the final deliverables.

2: Explore the data and create KPIs to track key indicators and identify meaningful relationships.

3: Draft the narrative arc of the analysis, using the 3 C’s: Context, Challenge, and Conclusion.

4: Create the final deliverables for the stakeholder (a multi-slide Story Points presentation).

Before we start constructing our story let’s understand the project brief.

Imagine you are an analyst working for the Super Store Corporation. Your boss Sylvia is the VP of Sales and she is conducting a review of the company discount policy. As research for her review, she has asked you to investigate if there is a relationship between discounted sales and profit, and how much the company is profiting or losing based on discounted sales. She will have only 5–7 minutes to give a story points presentation to board members, and she knows that her board members prefer quickly hearing only the top-level facts.


Ok, we are good to go…

1. Design checklist (Who? What? Why? How?)

Image for post

The design checklist is meant to be a reminder of some important considerations for your data storytelling. It contains 4 topics to be answered, who, what, why, and how.

So let’s answer the questions for our project.

#storytelling #data analysis

Creating Powerful Data Storytelling
1.40 GEEK