Visualization Best Practices for Data Scientists
Disclaimer: The ideas presented in this article are from the book: Story Telling With Data by Cole Nussbaumer Knaflic. To preserve the original message of the author, the visualizations presented in the book are also directly from the book.
When I was reading and enjoying the book, I thought it would be really cool to share my most key takeaways with the rest of the data science community. The book highlights the best practices for communicating effectively with data. It’s has been almost two months having all these tips drafted in my laptop, but thankfully now I realized it could help some data scientists looking to deepens their visualizations and communication skills.
In your opinion, how good are the following visuals? Hold on for few seconds and see what they are all lacking.
What would say about these visuals?
The main best practices for making effective visualizations and communicating the data are:
Understand the context: The need of communicating the data start with the reason to communicate. The author insisted on asking yourself questions like To whom are you communicating? What do you want your audience to know or do? How can you use data to help make your point?
Choose an appropriate visual display: There are so many visuals that we need to choose from when communicating data. But this can be a huge problem to our goal if we don’t understand the uses cases of each visual. The key advice is to choose a simple visual that our audience can understand. Visuals such as pie chart, donut chart, and 3D aren’t your best friend regarding effective communication.
Choosing effective visuals.
- Eliminate clutter: every time the audience see the visual, they start using their brain to understand it and in reality, everyone may interpret the visuals differently. To help the audience, we better has to remove any element that may take their cognitive power. We will achieve this by removing any element which doesn’t add value to the visual.
Before & After removing the non-useful elements.
- Focus attention where you want it: The audience read the visual in a zigzag way (Z). To convey information effectively, we need to keep track of our focus. We can control the attention by minding the size and the color that we use. Ask yourself this question: where are your eye drawn? This can help to put ourselves in the position of the audience thus conveying information effectively.
Making the best use of the color. What does this mean to your visualizations?
- Think like a designer: By using design concepts, we will highlight important stuff, eliminate distractions, use color strategically, remove unnecessary complexities and we will understand that white space is important.
Thinking like a designer means using every single element in visuals mindfully.
Tell a story: Like an exciting movie, the story starts with the magic opening, progress with attention-grabbing plays, and end with joy and a call to action. As the author articulated, the key rules to a great story are to talk about the subject of interest simply and to understand that the story is for our audience. It’s not for us.
“No longer will you just show data. Rather, you will tell a story with data.”
While these are what I could share with you, reading the entire book has more than that. Check the book blog storytellingwithdata.com.
Big thanks to the author for the amazing storytelling book, and you for reading this article.