This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Tech Community.
In the September 2015 issue of the Communications of the ACM magazine, there is an article on the Automated Education and the Professional - highly recommended reading. It feeds in nicely to a journey in learning Data Science.
That article covers the conflict between traditional college-degree education and the newer competency-based learning systems. I prefer both - a degree is extremely useful not just for landing a good job, but more importantly for a well-rounded education as a human being. As we think about learning Data Science, however, this question becomes quite important.
The traditional (albeit a brief one since the profession is relatively new) route of a professional Data Scientist is a degree in higher math and programming, with a focus on newer data processing technologies. This is complemented with a lot of experience, and self-learning using various methods of the latest programming and processing languages, along with Machine Learning, data visualization, and more.
In many cases, folks are learning as they go. Perhaps you have a few college courses in statistics and higher math, but you need a lot more. Happily, there are a lot of ways to learn the information you need to be a practical Data Scientist or Data Engineer, at little or even no cost. But before we dive in to the details, you need to find out how what you know now, and then how you learn new things.
Hubert Dreyfus, mentioned in the article above, finds that there are various "levels" of learned skill. He divides them up into the following:
- Advanced Beginner
Each of these levels determines what you need to know and how you learn. For instance, when you're brand-new to a skill, you learn by following rules. When you master that skill, you rarely rely on the rules. So in the first case you'll need to learn the rules, and later you'll learn the theory behind them.
To apply this to how you can learn Data Science, you'll find that you're in different places with the various skills you need. For instance, perhaps you are quite familiar with a Relational Database Management System, and a little less familiar with the NoSQL variants, and fairly new to Predictive Analytics and Machine Learning. Even within those areas you may have something I'm quite good at and something you've done less often.
So your first step will be to find out the skills and areas you need to know about Data Science, and then document what you do and don't know about each of those areas. We'll cover those areas in other articles here, but for now, you know this step is something you need to do.
Next, you need to understand just how you learn. There are a few places you can go online to find out how you learn, and I'd recommend you do that. "Know Thyself".
There are lots of ways to learn, and as it turns out, you probably learn different things in different ways.
Here are a few ways of learning things:
- Visual - Seeing a thing explained graphically
- Reading - Reading about a thing
- Experiential - Doing something to learn it
- Audial - Hearing someone talk about a thing
- Examples - Seeing a completed thing, and reverse-engineering how it was accomplished
Personally, I find myself mostly gravitating towards the Example-based learning style, but in fact the route that works best for me is to combine as many learning styles as I can.
Armed with the information above, you now have the ability to design a plan that covers what you know, and what I need to know. Your plan will look different than someone else, since you'll know things others don't, and you won't know things they do. You'll create your own plan as we move along. The articles on this Community will help you put this training into practical use using solutions as an example.
Before you begin - take this course: https://www.coursera.org/learn/learning-how-to-learn/
In a hurry? I have an article where I describe how I learn things quickly - you can read that here.