Historically, we could analyze data if it’s been made available to us as part of our business. However, we were forced to look at that data retrospectively and ask, What Happened? Only then would we be able to build a manual process around the analysis, make decisions based on that information, and finally act on it.
Somewhere along the line our curiosity got the best of us and we started asking, “Why?” Why did this happen in our data? But we resorted to the same process to answer our questions. There is nothing scientific about this. Don’t get me wrong, the data is great (and probably still useful), but it still requires human intervention to make the final decision. When we introduce the scientific method to our process, we create a repeatable experiment that not only gives you those answers, but automates the decision-making process.
Introducing Data Science and Predictive Analytics with Microsoft.
Data science is the statistical principals behind predictive analytics. Predictive analytics asks, “What’s going to happen next?” And then you get the answer. Real-time analytics sits on top of predictive analytics to support the automated decision-making process. We take new data coming in, such as an action/decision in real-time. Then we add intelligence components, like Cortana or Siri, on top of that so that we can interact with them confidently and gracefully in our day-to-day. Cortana is a perfect example of interacting with machine learning.
Now the question becomes, “How do we incorporate some of these experiments in our business so that we can leverage some of the answers and support data science gives us?” Well, you need data. So, we should always be interested in leveraging new data sources as well as existing data sources to create new relationships and understandings. As data scientists, we should be obsessing about data, relationships, and the de-identification process surrounding what matters.
Data science doesn’t happen in a vacuum.
It’s part of a larger business incentive, so the data needs to be presented in a way that the rest of your organization can understand. The goal is to display your evidence so you can solidify and explain the back story to your management team. When we can accomplish this, we can show how we’ve improved real-time processes like assembly line decisions or fraud prevention. Then we can effectively back up your intuition with data-driven analytic decisions.
Microsoft Data Science Professional Program
Edx.com: Data Science: 203.1x Data Science Essentials