As seen in Charleston Business Magazine, by John Lewis
Data science, the process used to develop analytics from the big data to use for research and development, is an inescapable topic these days because it is still developing. There are frequent revelations about its applications for businesses of all kinds. Many of the topics grouped under data science include classic statistics and data analysis that have been used for years, but with improvements in the computer fields of machine learning and artificial intelligence, the capabilities of data science now offer a greater range of analysis and forecasts.
Customers are telling businesses what they want and do not want through simple interactions, creating a wealth of data that can improve consumer satisfaction, produce product innovations, and increase efficiencies. The ever-expanding breadth of available data enables business strategists to plan as never before. Many companies have large amounts of underutilized consumer data such as purchase history, complaint calls, repair requests, and the like. There is also a wealth of information from internal processes of a company, which includes diagnostic information from industrial machines, error rates in order fulfillment, delays from individual suppliers, and more. This raw data has valuable information embedded within it that can help determine if, for example, there are products that consistently fail after a certain number of usage hours. The data could pinpoint whether individual suppliers are causing customer loss when they are late with deliveries or if a customer is probably ready for the next phase of a product to compliment a recent purchase.
Data science, until recently, was more associated with “big data,” which is comprised of the countless incoming bits of information gleaned from electronic devices, internet activities, credit card swipes, and other sources that business use to develop insights into customer usage patterns and operation effectiveness. It used to be a laborious, expensive task to even just house data created by business processes. Now, for example, businesses can track how long a person stood in front of a display in a store and how many times they returned to the display. A client’s visit to websites generates similar data. This automated generation of data creates an avalanche of information to be stored and studied. Companies now have very efficient methods to gather and store this information in onsite computers or in the cloud.
Data science ensures that the correct conclusions are assembled, enabling a business to make the most of the data itself. It is tempting to simply mine the data and jump to premature conclusions, but be wary − it is easy to grasp at false conclusions if only a small portion of a larger data set is analyzed.
There are established and well-defined methods for the application of data science to business problems. Some well-known proponents of these techniques are Google, Amazon, and Facebook. We are all familiar with Amazon’s recommendations based on a shopper’s previous purchases. Just five years ago, only large organizations like Amazon could create such “recommender engines” that filter data used to predict future customer activity. Today, what was once a highly arcane art is now a known business process with low-cost and easy-to-use recommender systems available to integrate into the websites of business of all sizes.
Data science, when properly applied, will likely improve business outcomes, and you likely already have the data you need to begin an analysis. Companies such as Dataiku, DataRobot, Google, and Microsoft combine the mathematics, computer science, and data management tools comprising data sciences into single packages, eliminating the need for experts in each of these topics. The tools create a workbench for examining ideas and pulling actionable information out of the data mountains. Most of these tools do require a basic understanding of statistics and familiarity with other instruments like Microsoft Excel.
After realizing the benefit to your business through data science, begin building on that knowledge through educational and professional development opportunities led by trained data scientists. Your competitors are already doing it and moving ahead because of it. There is no longer a choice about employing data science in order to build on a profitable future.
John Lewis is a visiting assistant professor of finance at The Citadel’s Baker School of Business. He has worked at the Organization for Economic Co-operation and Development, as a quantitative portfolio manager in London, Boston and New York, in venture capital for early stage companies in Paris and as a professor of finance at universities in Paris, London and Charleston over the past 37 years. His current research interests include quantitative finance, venture capital, machine learning and high-performance computing.