Data Science is to computer science what computer science is to electrical engineering. Back in the early days of computer science, electrical engineering was the dominant discipline. Electrical engineering defined the platforms and interfaces that dominated most developing industries. Over time, computer science relegated electrical engineering to commoditized objects of hardware, which were sewn together and operated by software systems. Computer science has radically transformed the world over the last generation, and enabled computational scale at unprecedented levels. We are entering a similar transition from computer science to data science, as data science starts to define the systems that enable next generation capability and business value.
Data science blends a variety of general techniques from applied mathematics, statistics, and computer science that are rapidly transforming our information processing capabilities and, therefore, our everyday experience. While no categories are pure, the questions that a data scientist asks and the technologies a data scientist wields are somewhat different than that of a computer scientist or even a statistician. Like a statistician, a data scientist is interested in patterns in data sets, but also like a computer scientist a data scientist is interested in practical cost and limit considerations of computing real and valuable results at scale. It is this mixing of disciplines and capabilities that enables data science to be a transformative technology.
As an example, a data scientist might ask: “how can we build, in a practical way, a valuable predictive model out of continuous streams of data which are increasing in volume and/or velocity by an order of magnitude every year or less?” This question may lead the data scientist to think about certain kinds of statistical models that are appropriate for streams rather than finite sets. Then, like a statistician, the data scientist must carefully analyze the real problem and the kind of data available on which to build and apply models that generate insights. Finally, like a computer scientist, the data scientist must help choose the right algorithms and consider the appropriate information-processing infrastructure to allow the whole concept to be realized and sustained economically. Very practical applications that relate to questions of this type are emerging in areas ranging from the Internet of Things (IoT), pharmaceuticals, insurance, and other financial services.
Following the impact that first electrical engineering and then computer science has had in all industries and aspects of our lives, data science is just beginning to transform our economy and society. Like computer science in the 1970s, today many data science techniques have been deeply explored and developed in research or focused industrial areas like computational advertising, cyber-security, and high frequency stock market trading. For the rest of the world, we are very much at the dawn of the data science era similar to the dawn of general computing era in the 80s. What sets this transition apart, however is as a collection of techniques and technologies, data science is capable of being even more transformative than any of it’s technical predecessors.
Written by Stu Bailey