More than 10 years ago we were colleagues, studying for a Communications degree that we hoped would get us a job in a PR department somewhere. And it did, but today we’re both preachers of a different type of communication – data visualization.
Rebeca Pop, the founder of Vizlogue, sat down with me to talk about how we should communicate effectively with data and what data visualization pitfalls we should avoid in data storytelling.
Rebeca, what was the moment when you started paying attention to data visualization? What made it interesting to you?
That’s a great question! To answer it, I’ll take a few steps back in time. I moved to the US in 2010 to pursue a Master’s degree. Almost half of the classes that I took were in research, analysis, and statistics.
Upon graduation, I started working in media analytics. The more I was expanding my data analysis skills and working with other people who were doing the same, the more my communication background started kicking back in.
In other words, I started realizing that the data analyst’s role was somewhat misunderstood. An analyst’s role should not be limited to analyzing data and generating a report. Not at all. Any analyst’s ultimate goal is to inform business decisions and to communicate with data. As soon as I identified this knowledge gap, I started reading book after book, and practicing my data visualization and storytelling skills assiduously.
You have been teaching Data Visualization and Storytelling at the University of Chicago and at DePaul University for the past 3 years. What is one of the most creative projects your students did for your course?
Since I started teaching data visualization and storytelling, I’ve seen many, many remarkable projects. What I’ve noticed consistently is that the key element that makes a difference between a good and an exceptional project is not necessarily the student’s background or prior knowledge. Rather, it is the level of passion that the student has for the topic and data being analyzed and visualized. A recent project that stood out to me was a visualization of TED Talks.
Other projects that stood out over time were primarily related to visualizing personal data or, as we call it in the field of data visualization, quantified-self projects. Topics ranged from phone usage, to sport performance and music preferences.
Tell me about Vizlogue. What was the motivation for you to transform your love for data viz into a business?
Whether I am teaching at the university or conducting a workshop for an organization, my mission and underlying motivation remains to help people communicate more effectively with data.
The reason I decided to make my passion a business is twofold. First, I realized that there was an even bigger gap in knowledge than I originally anticipated. Secondly, I started being contacted directly by some organizations, asking if I offered private training sessions.
That’s when it occurred to me that I could transform my passion into a business.
How can data viz help companies communicate better? Do you see a change in how companies perceive the value of data visualization?
To answer your first question, there has been a hype around data in the last few years. In the business world, we hear the term “data-driven” so often that it almost lost its meaning. I don’t think that most of us truly want to make data-driven decisions. Because one data set can tell many different stories. Ultimately, I think we want to make data-informed, but audience-driven decisions. And that’s where data visualization and storytelling come into play.
Regarding the second question – absolutely. I see companies and organizations being hungry for data visualization and data storytelling knowledge. Organizations realize there is a gap in the formal training that employees received. Now it’s time to fill that gap via additional education.
We both agree that in order to experience the real value of this type of communication, you must also avoid the pitfalls.
What are the most common data visualization pitfalls that you’re seeing in your everyday practice?
Definitely. I like to think of data visualization as not being just about visualizing data per se, but also about pausing and evaluating the data very systematically before analyzing and visualizing it.
I don’t know if I can pick one most common data pitfall. A few that stand out and seem to happen recurrently are:
- Not checking and validating the source of the data;
- Not humanizing the data;
- Not checking and validating the metrics and data collection methodology.
I should also specify… I think that it is very easy to fall into data visualization pitfalls. We all do. I still do sometimes. That’s why I always recommend not to underestimate the amount of time one needs to spend understanding the data. It truly is an essential step in any data visualization process.
If you had to choose, what would you say it’s the one data viz pitfall that does the most harm?
I don’t think I can choose one, as it depends on the impact that the pitfall ends up having. But I would say this – I have seen many data pitfalls in COVID-19 visualizations, and that is very concerning. Whenever we work with highly sensitive data, the effects of pitfalls can be disastrous.
One pitfall that we see is that businesses don’t always make sure their data is accurate. What are some solutions for detecting data inaccuracies?
I think this goes back to what I mentioned earlier in regards to data understanding. It is key to allocate sufficient time to understand the data before jumping into creating visualizations. I typically recommend writing down a list of questions to ask ourselves as we examine the data. I’ll give some examples of such questions:
- Where does the data come from?
- How was the data collected?
- Why were these specific metrics used?
- Is there any missing data?
- Are there any outliers?
- Is it likely that the data is wrong?
I should also note that I don’t think this list of questions should be static. We should be constantly adding to it, as we become aware of additional pitfalls.
“Well, the data speaks for itself.” - data viz pitfall or not?
Definitely a pitfall. The data never “speaks” for itself. We make the data “speak,” based on our own knowledge, background and biases. Let’s put it this way – any data can speak multiple languages. It’s on us to pick the right language.
Thank you for taking the time for this interview, Rebeca.
The last question is more personal. Here it goes: what type of data you couldn't live without?
Hmm… I’ll answer this question in the context of current events. Right now, I couldn’t live without COVID-19 data (as accurate or inaccurate as it might be). This informs many decisions that I am making on a day-to-day basis.