Importance of Communication and Visualization Skills in Data Science

Last week, I lead the second session of the Finland-Uganda Data Science Meetups. The topic for this session was data import, cleaning and pre-processing, and visualizations.

In the session, we welcomed senior data scientist, Jukka Toivanen, who shared with us the importance of good communication and visualization skills in data science. With the image below, he outlined to us the three spheres that make up the major part a data scientist’s work.

In those three areas, he pointed out that there is a tendency of data scientists to focus on the comfort zones, i.e., the blue and purple spheres. With the red sphere often not being focused on so much.

“Effective communication is undervalued but important” – Jukka Toivanen

But communication is important and it often means the use of good visualizations, as they can have remarkable effect in spotting patterns and conveying them effectively to other people.

I also recently got to learn the importance of this skill in an interview with a company for a data scientist position. The company gave me a task that involved developing a predictive model on a dataset, with the additional task of presenting key findings to a CEO. Guidelines were that the presentation should  be maximum 10 slides, be visually appealing and should be top notch. My performance on the task would be based (not solely) on my presentation, i.e., communication skills.

Although, I have undertaken several presentations in my career, this task did challenge me. As an audience, executives are a special group. They have limited time and want to hear meaningful results quickly. Thus, my first two slides had to already catch their interest. In addition, with visualizations, I had to communicate two to three important findings, insights and action points that had the most business value. To do this well, required learning as much about my audience’s business as I could and their potential business model. This paid off in the end as I was commended on my impressive presentation.

Often though, good visual presentations is not something that is emphasized in many data science school programs, let alone computer science programs. But it is important to learn how to effectively communicate and visually connect the technical approaches and results to the business value. This skill I believe will take data scientists a long way in creating impact.

Also published on Medium.