Marc Soares is the Director of Technology & Solutions at Dandelion. He is a constant advocate for effective data visualization and storytelling. His preferred tools are Tableau and Google Data Studio. A lot of his work is featured in Makeover Monday, a social data project, and in the Data Studio Report Gallery.
Marc’s message revolves around simplicity in data visualization. This idea caught our attention, so we wanted to dig deeper. Luckily for us, Marc agreed to answer our questions.
Marc, what do you think is the main purpose of data visualization?
The main purpose of data visualization is to improve our ability to understand and communicate with data. Visual representations make it easier for our brains to see patterns and make comparisons. Data visualization can be used to identify relationships, support conclusions, and encourage better decisions.
When talking about your work, you mention that you aim for simplicity. At the end of your projects, how do you make sure you’ve kept it simple?
The French author Antoine de Saint-Exupéry wrote that “perfection is finally attained, not when there is no longer anything to add, but when there is no longer anything to take away.”
In this same spirit, I aim for simplicity in my visualizations by seeking to remove anything that isn’t essential. I’ll often experiment with removing an element or making it more subtle to see if it improves the viz. Before completing a visualization, I carefully consider every component—charts, text, lines, labels, colours—to ensure that each positively contributes towards the final result.
How can you make complex data fit in an easy-to-grasp representation?
Visualizing data with multiple dimensions and complex relationships can be a challenge. I think the best approach is to ensure there’s a clear message and that the chosen type of visualization supports that message.
It can often be helpful to break down complex data into smaller pieces and focus on specific aspects, rather than trying to show everything at once. Using a storytelling approach with text, annotations, and even animation, can also help to break down the complexity and guide the reader through the visualization.
Are there mistakes in data visualization that practitioners in the field should be aware of? For example, should all graphs start at 0?
There are definitely common mistakes with charts and graphs that practitioners of data visualization should be aware of. Bar charts in particular should have a baseline of zero, otherwise the lengths of the bars can be misleading. Axis scales and intervals should be even and consistent to enable fair comparisons.
It’s also a good idea to avoid plotting different metrics using dual axes on the same chart, as this can possibly highlight spurious correlations.
(…) when designing reports or dashboards for business stakeholders, the needs of a marketing analyst are very different from that of a Chief Marketing Officer.
There’s a trap many of us fall into when working on new projects. We forget who the audience is. How can we be more audience-aware with data viz projects?
I think it is essential to define the audience at the very beginning of a visualization project. The intended audience will influence almost every aspect of the visualization, including what data is included, how it’s described, and how it’s represented.
The same data may need to be visualized differently for different audiences. For example, when designing reports or dashboards for business stakeholders, the needs of a marketing analyst are very different from that of a Chief Marketing Officer. By defining the audience and identifying their requirements up front, you can save time and effort and ensure better results.
One of Marc’s projects for #MakeoverMonday, a beautiful tribute to Anthony Bourdain
Do you feel that sometimes the focus in data visualization is more on the design, and not on the actual data? Is this a good or a bad thing?
Yes, I have seen countless examples of visualizations that focus on the design, rather than the data. The design choices, while visually appealing, often make it difficult to understand the data. Given the purpose of data visualization I mentioned earlier, I don’t think it’s a good thing to prioritize form over function in data viz.
However, if the author is actually intending to create data visualization as art, then the same principles don’t apply. Data art can be purely aesthetic, and if that’s the goal, data can generate beautiful designs.
Have you encountered situations when data viz was used when not actually needed? When would an Excel / Google Sheets table be better than transforming the data into a visualization?
Tables have the benefit of density; a table can display a lot of data across many different dimensions and metrics all at once. A data table can also be very dynamic, providing the ability to search, sort, filter, and quickly create calculations and aggregations. So, tables can be useful for exploration and analysis.
However, visualization still provides a complementary function for seeing patterns and making comparisons. In many tools, simple visualizations like bars, sparklines, and heatmaps can be overlaid or embedded into tables. You can also combine tables and charts into the same report or dashboard to provide the best of both worlds.
Marc, you once mentioned that you fell in love with Google Data Studio because of the ease with which it lets you connect the data, create visualizations, and share them with others.
You also mentioned that there are some powerful features missing. What features do you think should be added?
That was over 3 years ago! Data Studio has come a long way since then. At the time, I was referring to features for data transformation and analysis, such as data blending and advanced calculations which have since been added.
Today, I think Data Studio has more than enough functionality for most people. And the native integration with BigQuery and Google Cloud Platform provides all the data transformation and analysis power you could ever want.
I would still like to see some improvement and refinement to the built-in charts; certain style options are not available or unable to be customized. It would also be useful to have better functionality for organizing and managing reports, for example folders and user management. I’m also looking forward to seeing further development of the geospatial mapping features that were recently added.
Speaking of Data Studio, how much do tools matter when the goal is simplicity in data visualization?
I think tools can play a significant role in promoting good data visualization practices, including simplicity. When data viz principles are incorporated into default chart settings and templates, it makes good data visualization more accessible and attainable for the average user.
Data Studio does a good job of keeping it simple in the default styling of charts and tables.
Thank you for sharing your knowledge and your time with us, Marc! Here’s my last curiosity: what’s the most powerful story you encountered while creating data visualizations?
The most powerful data viz stories that I’ve encountered have been related to the COVID-19 pandemic over the past year. Data visualization has been a critical tool for governments, public health agencies, and media outlets to understand and communicate rapidly changing conditions around the world.
Early in the pandemic, I contributed to an effort called HowsMyFlattening, a collaborative effort between public health experts, medical professionals, and data practitioners to track our local response to COVID-19 and help the public better understand what was happening.
The concept of “flattening the curve” is by nature a visualization-driven message, one that is literally a matter of life and death. I think the charts and reports that I have created about COVID-19 have been the most meaningful data visualization work I’ve done so far.