"It is important [...] to also understand what comes before and after creating visualizations.” - Isabel Meirelles
Let's be honest. Visualizing data is not a straightforward process. Sometimes, you need to skip a step, or move at a faster pace...or simply take a few steps back.
Regardless of how you'd need to adjust the 8 steps below to fit your specific use case, one thing remains certain: data visualization means much more than simply designing charts. Designing charts is just the middle point. Understanding what comes before and after visualizing data is equally important.
In this step-by-step data visualization guide, we'll explore the data visualization process holistically. I'll illustrate the 8 stages of visualizing data by using a graph that I created in the past.
The 8 Stages of Visualizing Data:
Understand your audience
Understand your data
Data collection
Data transformation
Find the story
Sketch
Create the visualization in a tool
Receive feedback and edit
Do you want to learn more about the process of visualizing data? Check out our workshops, which dive deeper into the visualization process and incorporate hands-on exercises.
Visualization: Amazon's Profit Margins
1. UNDERSTAND YOUR AUDIENCE
Typically, I spend a lot of time understanding my audience before exploring a data set. Why? Because the audience should guide what metrics or data points I am going to use. This is important when working with small data sets, and even more important when working with big data.
For the visualization above, my audience was broad -- the average consumer. So my goal was to visualize metrics in a way that would be easy to understand regardless of the person's level of financial literacy.
Once I was done evaluating my audience, I moved on to the next step: data understanding.
2. UNDERSTAND YOUR DATA
The MakeoverMonday team provided the data set as part of their weekly data visualization challenge. The original source was macrotrends.net. The original visualization came from this Vox article.
The data set included 3 columns: quarters, Amazon revenue, and Amazon net income. I noticed after looking at the data that the story presented in the original Vox article was indeed true: Despite seeing its revenue skyrocket, Amazon's profit remained low.
While the story and the data were insightful as originally presented, I was left wondering if the metrics could be more intuitive. Instead of revenue and net income, profit margin might be more relevant for the average person.
Also, the original visualization in the Vox article was missing some context. By adding context, I could answer questions such as: "Is Amazon's profit margin unique?" or "Does Amazon's profit margin follow the same trend as other large tech companies?"
3. DATA COLLECTION
Typically, data collection comes before data understanding. This visualization was an exception. Although I received data that had already been collected, I wanted to add more context. So I went back to what would typically be step 1: data collection.
I downloaded two additional data sets from macrotrends.net: 1) revenue and net profit for Google and 2) revenue and net profit for Microsoft, all by quarter. Then, I merged the original file with the two new ones.
Here's how the new data set looked like.
4. DATA TRANSFORMATION
Given that I could not find data for the metric that I wanted to visualize (profit margin), I created it myself.
It took a few simple Excel calculations and I went from the previous version of the data to the spreadsheet below.
5. FIND THE STORY
When I work with large data sets, I typically spend a good amount of time exploring the data. Exploring the data helps me in the process of finding the story. In this case, the data set was small so I didn't spend a ton of time exploring it.
How did I find the story? Digging into the data, I saw that Amazon's profit margin was fairly consistent. Google's and Microsoft's profit margins fluctuated more over time.
Done! I found the story in the data.
Next, I grabbed a pen and a piece of paper.
6. SKETCH
Data sketching is my favorite part of any data visualization process. It allows me to immerse myself into the data and envision my visualization without the technical barriers of a tool (any tool!).
I knew I wanted to show how profit margins for Google, Amazon, and Microsoft trended over time. I also knew that I was a fan of clean, clear, minimalist design.
So, I sketched, and sketched, and sketched.... You can see below a few of my sketches.
Note: I may not have the most beautiful handwriting but I hope these sketches help you understand my thought process.
7. CREATE THE VISUALIZATION IN A TOOL
Once I looked at my sketches, I decided that it didn't make sense to show all three metrics (revenue, net income, profit margin), as they were somewhat redundant. Instead, I wanted to focus on net profit.
In terms of position on the page, I thought it would make more sense to show Amazon separately from Google and Microsoft, as the scale was so different. Amazon's story would have been partially lost had I put all three trend lines in one chart.
Here's how my first versions of the visualization looked like in Tableau.
8. RECEIVE FEEDBACK AND EDIT
Given that the audience for my visualization was broad (the average consumer), I asked my husband for feedback. Receiving feedback is such a critical aspect of data visualization. It can help you see a totally different perspective.
Based on his feedback, I made a few small design adjustments, primarily related to the use of color and text.
This was the final visualization.
FINAL THOUGHTS
Going back to the beginning of this article, I hope that these 8 steps helped you see why what comes before visualization design is so important.
Now, you might be left wondering what comes after data visualization. My answer: storytelling. If you have to share your visualization via e-mail or, in particular, if you are planning to present it in-person or online, then storytelling is the next step.
I'll leave you with a collage of the process that I described in this post.
It is very informative and good explanation, been helpful! Thank you for sharing.
I think the image right at the beginning is the most appropriate. Building out informative and quality visualisations certainly isn't a linear process!