Innovative Education: Educating the next generation of data scientists

By Deepinder Dhingra and Meena Anantha Padmanabhan

In today’s turbulent and competitive business environment, analytics can be a powerful decision-making tool. In a world of signal deluge, even creative organizations need a mindset of measurement and improvement to optimize their decision-cycle times. As a result, increasing numbers of companies are moving away from intuition and gut-based decision making and turning toward data-driven decision-making.

In their endeavor to become data-driven, organizations need a systematic framework to think about the different types of analytics needed to create insights and help make better decisions.

The following framework describes the different kinds of analytics needed:

• Descriptive analytics answers the questions, “What happened in the business?” It is looking at data and information to describe the current business situation in a way that trends, patterns and exceptions become apparent.
• Inquisitive analytics answers the question, “Why is something happening in the business?” It is the study of data to validate/reject business hypotheses.
• Predictive analytics answers the question, “What is likely to happen in the future?” It is data modeling to determine future possibilities.
• Prescriptive analytics is the combination of the above to provide answers to the “So what?” and the “Now what?” questions. For example, what should I do to retain my key customers? How do I improve my supply chain to enhance service levels while reducing my costs?

This is not a linear journey; all four types of analytics can and should be combined in the right mix. Organizations that focus on only one aspect of what is called the “DIPP” framework will fail to generate the right insights and recommendations.

Figure 1: Analytical framework – data in, insights out.

However, the creation of insights alone is not sufficient. Going forward, companies will compete not so much on the creation of insights, but rather on the consumption of insights. Consumption entails communicating insights, implementing insights, measuring, incentivizing and developing cognitive repairs. It refers to the planned, ongoing use of a set of interlocking business practices and competencies that collectively delivery superior value from analytics insights. Enabling consumption will need an appreciation of behavioral sciences and how organizations and human beings absorb new and often counter-intuitive insights and process them to adjust their cognitive machines to make decisions.

Thus, the future of analytics will not just be based on applied math, business and technology, as it is today. The future will witness the notion of analytics evolving to decision sciences encompassing math + business + technology + behavioral sciences.

The past, present and future of analytics can be summed up as follows:

    Yesterday

• Business + technology allowed us to simply automate.

    Today

• Math + business allows us to present more cogent arguments at the boardroom.
• Math + technology allows us to operate proactively with anticipation.
• Math + business + technology allows us to execute better.

    Tomorrow

• Math + business + technology + behavioral sciences will let us develop nudges (cognitive repairs) against biases that we as human beings are gifted with.

Decision Sciences & Analytics Education

Based on the foregoing, a formal educational framework for decision sciences should enable creation and consumption of analytics and inculcate a culture of data-driven decision-making by imparting the necessary knowledge, skills and values.

One of the shortcomings of existing educational programs is that they tend to focus only on analytics techniques, applications, technologies and data; essentially, the data science aspect of decision sciences. Ideally, more holistic educational programs will recognize the key imperatives and challenges for making decision sciences successful in an organizational context, and will create programs for different roles that come together to enable data-driven decision-making.

Figure 2: Creating insights and the consumption of insights is not a linear journey for businesses.

We believe that a new army will emerge at the forefront of analytical competition, and that formal education programs must work toward developing curriculum for these varied roles:

•     Soldiers: These are typically analysts and data scientists who work on solving business problems and generate and communicate findings and insights. They will come from varied backgrounds including engineering, computer science, economics, math, statistics, business and other quantitatively oriented fields. They will need to develop skills in analytics techniques, data, technologies and applications along with a combination of consultative, first-principles-based thinking for problem definition and hypothesis-based approaches. Also, they will need to bring together right- and left-brained thinking to balance the rigor of science with the creativity that business requires.

Principles of design, usability and visualization will be key to making them successful both in the creation of insights and consumption enablement.
•     Captains: These are middle managers driving analytical initiatives. They are usually quantitatively oriented professionals with experience in both analytics and functional roles such as marketing, risk, supply chain, etc. Agile and iterative project management skills required for analytics will be essential to adapt to the dynamism of the business problem environment and to manage new processes that cut across functional boundaries. They will also need to develop knowledge management frameworks leveraging insights from across verticals and domains to drive innovation in addition to effectiveness and efficiency. The ability to work with geographically dispersed teams will also be a key requirement in a world of “glo-calization” where bringing local perspectives will be essential to success in developing and emerging markets.

These captains will play a pivotal role in consumption of analytics since they need to simultaneously play the role of explaining the science behind the analytics to the business user and translate findings into insights and recommendations. Thus, understanding of behavioral sciences and the role of cognitive biases in decision-making is a key function.

Figure 3: Math + business + technology + behavioral sciences = future of analytics/decision sciences.

•     Generals: These are senior management folks with a strong vision and passion for data-driven decision-making. They will come from diverse backgrounds and will be organizational leaders with deep business acumen and an appreciation of data-based insights. They will need to develop skills in creating and running cross-functional analytics councils, shared services eco-systems, governance models for analytics organizations, analytics roadmaps, etc.

Change management and the ability to bridge the gap between creation and consumption of insights will be their key focus.
Data-driven decision-making is a journey and without the right talent across organizational levels, the benefits of decision sciences cannot be truly realized. Decision sciences education is in a nascent stage and needs to evolve to a holistic approach. What is needed is an interdisciplinary approach drawing its foundation from mathematics, business, technology and behavioral sciences.

Deepinder Dhingra is head of product development at Mu Sigma (www.mu-sigma.com) where he is responsible for recommending solutions that can scale and meet customer requirements in real time with the help of assets that are developed in Mu Sigma. He has more than 12 years of experience in consulting and sales roles. He previously worked for various enterprise software firms in the areas of business planning and business intelligence.

Meena Anantha Padmanabhan is an associate manager with Mu Sigma where she works with multiple Fortune 500 clients, across multiple verticals, solving high-impact business problems in the areas of marketing, risk and supply chain analytics.