Data Visualization for Decision Making - Navigating Uncharted Waters

Prof. Kavitha Ranganathan

In 2011, as a relatively new faculty member at IIMA, I was trying to figure out what courses might add value to our MBA participants. As a computer scientist who had worked primarily on algorithms and protocols for distributed computing, the highly technical nature of my training would mean I had to adapt fast to a completely different environment. Data Visualization was becoming a buzzword in many forums and I was as curious as any other person about what it entailed. With a little encouragement from my husband (also my colleague), taking up the challenge of a completely new topic snowballed into something I could never have anticipated!

Apart from a two-day workshop on Data Visualization I had attended in Chicago as a grad student many years ago (by Edward Tufte no less!), my exploration on the subject started with a blank slate. A few quick searches revealed that existing courses in Data Visualization fell into two broad categories. Some courses were offered by Computer Science departments and were highly technical in nature involving large amounts of programming and esoteric constructs. Others were part of Design School curriculums and dealt more with Infographics and emphasized design elements rather than data analysis. A course for our MBA students would need to include practical and relevant skills in creating graphs that made sense of data, and in gathering insights and communicating them effectively. I would need to design a new course from scratch. Data Visualization is interdisciplinary in nature - it draws on statistics, design and communication, and the new course would need to balance out all three aspects by incorporating elements from all three disciplines.

While the seminal work by Tufte provided an anchor for the theoretical underpinnings of the course, it was the work by Stephen Few that provided practical tips and actual business scenarios. After an intensive period of reading, learning, assimilating and preparing new material, a course began to emerge. To the credit of my colleagues at IIMA, I received nothing but encouragement to go ahead and offer the new course as an elective for our second year students. The first offering was a short 0.75 credit course (15 75-minute-long sessions) and it attracted 30 eager participants willing to invest in a foreign topic. Their feedback was very positive and by the next year, the demand for the course had shot up. I expanded the offering to a full credit course of 20 sessions; soon two offerings were needed, and a third one for the Postgraduate Programme for Executives. Today, the topic attracts a large number of students across programmes and the feedback from alumni on the usefulness of the course in the workplace has been very gratifying. The course has also spawned an executive education programme on Data Visualization which generates a lot of interest.

The course as currently taught at IIMA uses a mix of lecture sessions, in-class exercises, lab sessions and group critique sessions. Lecture sessions deal with, among other topics, frameworks for identifying the right graph for the right message, visual perception rules that can be leveraged for effective graphs, storytelling with data, and understanding the nuances of effective design. The value of these sessions is realized when participants apply the principles learned to specific business scenarios and cases. Lab sessions on tools like Processing (for interactive visualizations) and Tableau (for exploring data) add range to the participants' data-handling and data-shaping abilities.

With such a variety of software tools becoming available now, it might be mistakenly assumed that a data visual is something anybody can whip up rather quickly. While that statement is partly true, creating an effective visualization is a whole different story. Apart from understanding the theory behind visual perception, pre-attentive attributes, design elements, form and function and colour choices - just to name a few aspects, creating an effective visualization requires loads of practice. Thus, the participants are not only exposed to a large number of exemplars for a range of scenarios but are given a plethora of assignments and in-class exercises. One particular session that many students find very helpful is where their own creations are critiqued by the whole class. We do anonymize the submissions before critiquing them, however! The creators quickly learn that what had seemed perfect to them did not appear to be so when others tried to make sense of what was being communicated. This experience also makes them realize that visualization is not an end in itself, it is only a means to an end, and hence should always be created with a specific audience in mind. The course culminates with a group project where students are encouraged to apply their learnings to a topic of their choice. These projects have led to very creative and fun explorations.

The journey has been very gratifying and fulfilling. I have not only been able to equip many students with a very helpful and essential skill-set in today/s data-driven world but have also managed to learn a lot myself.

About The Author

Prof. Kavitha Ranganathan

Prof. Kavitha Ranganathan

University of Chicago, Ph.D. (Computer Science), December 2004
University of Chicago, M.S. (Computer Science), June 2001
Birla Institute of Technology and Science, Pilani, India, M.Sc. (Information Systems), June 1998