Making data science a team effort

Professor & Information Systems Department Chair, W. P. Carey School of Business at Arizona State University

Using higher education for player development


With the world’s best football teams facing off for the FIFA World Cup this summer, all eyes are on the soccer fields of Brazil. If you don’t follow this type of football (or maybe you even call it “soccer”), then the player positions may not be on the tip of your tongue. Take “strikers” for instance, the role most likely to score goals and win glory. They need a whole cast of characters in other roles, such as goalkeeper, midfielder and forward, to have a viable team. By orchestrating these roles, team-level strategies and tactics are carried out. The same can be said for data science.

Many make out the data scientist to be a Renaissance woman or man: a modern-day, corporate-recognized striker who can single-handedly elevate the organization’s analytics savvy. Jeff Bertolucci of InformationWeek refers to this ideal person as a data science “unicorn.” However, as in football’s version of the situation, a team can play the data science role better than any one individual. Each individual is probably best at one or maybe two roles, tops.

Accenture identifies the following data science roles: Systems architect, quantitative analyst, software engineer, visualization designer and business analyst. SmartPlanet’s “The Bulletin,” describes the roles of data architects, data visualizers, data change agents, data engineers/operators, data stewards and data virtualization/cloud specialists. There are also roles in machine learning, statistics, optimization, simulation and more

Success at data science can be a team-level effort with individuals combining their skills, knowledge and experience. Preparing students for corporate roles in data science means graduating more than just strikers; for different team configurations at different companies, there will be requirements for more or less of certain positions.

For example, one organization might have its data science team develop a new predictive model, and then throw it over the fence to the IT department to deploy. IT might deploy it by embedding the model into automated business processes. Then, who monitors the predictive model’s performance? Will that be an IT responsibility, or will IT throw it back over the fence to be caught by another player on the data-science side?

0829.Zhan-Michael-Shi-Feature-101Student.jpgFor educational institutions with programs in analytics and data science, the important role of industry advisory boards in helping to develop real-world curriculum is crucial. Also, there’s another important role for industry support: the one from vendor companies working in the trenches with clients on the most wicked of today’s data science problems. What they learn and the tool suites they evolve are critical for designing the right educational experiences to prepare students for all data science roles.

At the W. P. Carey School of Business at Arizona State University, we work very closely with our industry advisory boards to create curriculum for programs in business analytics. Our Master of Science in Business Analytics program matriculated its charter class in 2013-14. The program is a coordinated effort of the school’s highly ranked information systems and supply chain management departments. Entering students come from a variety of backgrounds and different areas of in-depth experience. This diversity, combined with the program’s curriculum, uniquely prepare each individual for data-science positions. We embrace the fact that different students have natural strengths they bring to the table for different data science team configurations.

This fall, a new undergraduate degree program in business data analytics will also begin. The undergraduate degree teaches skills and knowledge leading to an entry-level data science team player position. We are already seeing considerable interest from students wanting to combine the analytics major with economics, finance, information systems, supply chain management and marketing concentrations.

While our programs are new, they are undergoing significant and constant improvement. Companies like IBM have taken a growing interest, and partnered with us, in helping students learn all they can, using the most innovative of software suites, such as Cognos and SPSS, and methodologies, like Business Analytics Solutions Implementation Method. IBM has also helped with faculty training on the newest and best big data methods and tools. This level of support makes invigorating the classroom with the latest and greatest much more manageable.

Like both football and data science, developing viable higher-education programs and coursework in business analytics takes a team. We sincerely appreciate the support of our executive advisory board partners and companies like IBM that help keep us on the leading edge.

Training a world-class data scientist prepared for any industry, company or problem is difficult, if not impossible. Higher education is responding by preparing a next generation of data science team players who can fill that elusive one-size-fits-all shoes. Industry advisory boards are guiding higher education in designing programs that address the necessary skills. If you are a data scientist, get involved by joining one of these boards, and help to guide the next generation. Companies like IBM are stepping up with software and training support. If you are an analytics company, reach out to higher education institutes working on these new programs and offer your expertise, software support and faculty training.