Alpine Data Labs 2014 Predictions
Data science is permeating every facet of our daily lives - from our culture to our classrooms. Look for data science to make an even greater impact in 2014.
Guest blog by Joel Horwitz (Alpine Data Labs), Dec 24, 2013.
"Consumerization" of Advanced Analytics
More and more we're seeing machines replacing human decision-making. In 2014 we will see this trend accelerate. Take for example Google's self-driving cars or Amazon's drone delivery system that is in the works. Both of these projects are fueled by better algorithms or machine learning and are far less dependent on any physical constraints of the machine or environmental properties of the surroundings.
In a more near-term example, we expect to see more consumerization of these advanced analytics in everyday activities such as decreasing your energy bills by smarter thermostats (Nest), better living by predicting health based on food, activity, and stress levels (Fitbit), or how to find the right partner, career, pet, town, or simply fashion fit.
Today we aggregate findings from academic research, shared experiences, or marketing campaigns with utter disregard for what makes you unique. Starting next year, we'll see micro or even nano-personalization for all aspects of our daily life through advanced analytics at scale.
I for one look forward to when we stop talking about everything in terms of averages and instead take a more holistic approach to factoring in even more variables in our daily lives.
With barriers to entry slowly being reduced, high school students will learn data science fundamentals as part of their curriculum.
Data science principles such as, data mining, exploratory analysis, model selection, and validation criteria will be taught alongside physics, chemistry, biology, and many of the core sciences. It is highly unlikely that today any aspiring mind is answering "Quant" on their What I Want to Be When I Grow Up!" essay.
What is a Data Science curriculum and how is it different than statistics? Most people would cite programming, statistics, and business acumen at the heart of the core competencies. However, I disagree and see it more of a philosophy or language upon itself. For example, lets take chemistry - it is built upon the tenants of the fundamentals including math, critical writing and physics primarily. Similarly, Data Science is built on math, data interpretation, and experimentation. Furthermore, chemistry is the study of how atoms interact to form molecules, which interact to form larger organisms. Data Science is the practice of taking data elements, modeling, and constructing applications that interact and provide output information.
As Data Science moves beyond the constraints of programming, and we develop a common framework and terminology, it'll be far simpler to apply to any imaginable career path. I can recall a time in business when K-means, Naive Bayes, Random Forest, and the like were unknown. Soon, we will not be asking for any any old model to get the job done, but we'll have a more specific approach to get the best performance. We can already observe an uptick in interest over time in our own search behavior - see chart above or check Google Trends for "random forest", "deep learning".
Joel Horwitz is a Director of Product Marketing, Alpine Data Labs. He describes himself as "a passionate data guru educated in engineering, chemistry, and data science". He has MS in Nanotechnology from the U. of Washington and an International MBA from the U. of Pittsburgh. Previously he worked at AVAST Software, Datameer, and AVG Technologies.