
It’s often quoted that Data Science is the sexiest job of the 21st century. In that respect, to some degree, it makes sense why so many want to learn Data Science.
Data Science is a massive field, and often approaching it as an nonacademic can be extremely daunting. How do I know? Look at my LinkedIn profile…
I have gone from a Post-man to a Post-room assistant to interning as an Machine Learning Engineer, and to reassure you even more, before that I was playing Football at a professional level, and if we go back even further, at one stage I was poised to be a major Taekwondo fighter— I know, that’s a lot of changes, talk less of what I have not shared, but maybe I can go into that in a later post.
My greatest skill is my ability to learn. I’ve always known this. I have never believed I can not learn something, but I’ve never known why or how I done it. I have taken time to research and reflect on how I have learnt to get to where I am now and thought that it is only right I share the information that I have found out with my peers, so we can together accelerate the field of Data Science to unimaginable heights. Therefore, I have made it my responsibility to share how I learn Data Science with as many people as are interested to know, in order to accelerate to growth of Data Science.
The Process
There is a process to learning. Why is this important? Knowing that there is a process is important because processes describe how things are done. If we know how things are done then this will provide us with the potential to focus and to make the process better. If we can improve the process, we have a higher possibility for successful outcomes.
If you focus on the right processes and in the correct way, you can design your way to becoming successful. – I don’t know if this is a real quote by somebody else, but if it’s not I claim first dabs!
Ever heard the saying "fall in love with the process". Yeah, it’s pretty common, but we hardly think through what we say. I will prove it, fall in love with the next stranger you see. I said that to prove to you that it is very difficult to fall in love with something you don’t know, so if we want to fall in love with the process, we need to know the process. We know that mastery takes time. It requires planning and deliberate focus for a sustained period of time. With this in mind, we can begin…
In order to plan your Data Science career effectively, it is important to know where you want to go and then figure out how you want to get there. Here are some questions you may want to answer:
What are my Objectives?
Where do I want to go?
and to figure out how to get there, it is important to know where you currently are. You may want to ask yourself:
What skills do I have?
What skills do I need to reach my objective?
In answering these questions you’ve done 2 things that are vitally important to attain that new skill. First you have given yourself a sense of direction, and secondly, though it may take some research to know what skills you need if you are new to a field, you have given yourself a route to follow. Applying your energy to each of the skills that you are required to learn can be thought of as the motion driver to attaining that overall goal.
"To get to where you are going, you’ve got to know where you are" – Bob Proctor
3 Stages of Learning
There are 3 stages that we pass through when we acquire new skills and there are many skills required as a Data Scientist, from understanding these stages is what makes learning exciting. Why? Because when you review the stages, you can point to where you are, which will give you better clarity of what you need to do.
Cognitive Stage
Problem: In this stage, the problem we want to solve is understanding what we have to do.
It is very difficult to learn a topic without receiving prior knowledge about what you’re learning.
Many decide to go University, many prefer books, many prefer online courses. What do they all have in common? There is a someone, a teacher, that has amassed prior knowledge of the subject and they are passing on their experience to you – I have seen many people asking whether a PhD or a masters is required to learn Data Science and i’d say it’s subjective. You have to assess your objectives to make that decision.
Characteristics: You can expect to experience large gains in your performance with a lot of inconsistency. Instructions is necessary in this stage, and so is feedback.
An example of an instruction in relation to Data science could be to learn 2 algorithms per week and code them out from scratch. The feedback from this task can be given by testing your solution against working implementations to see if your scores are the same.
Goal: The goal of this section is to have information ingested and meaningfully organized in a transferable format that can be used in the next phase.

Associative Stage
Problem: The problem we are solving in this stage is that of transferring potential power into realized power.
We have all heard the saying that knowledge is power. In the rise of the information age, where there is tons of information that we have access to through one or two clicks, it turns out that knowledge isn’t exactly power. Knowledge is potential power, since knowing something doesn’t necessarily affect change.
We accumulated the know how of the skill we want to learn in the cognitive stage, amassing bags of potential power in the process. To realize that power we must act on it.
Practice, Practice, and more Practice!
I’ve been an athlete on two counts, as you see from my brief CV above, and one thing I can recall from this phase is the hours of repetition that go into it. I’d perform the same reverse kick for an hour, or when I was playing Football, i’d be taking free-kicks for hours to perfect my technique (If I can’t find a video of me, I may get my boots of the hanger if enough people request to see).

In relation to Data Science, this could be simply feature engineering. Want to be good? Take the knowledge you’ve learnt about feature engineering from the cognitive stage and deliberately apply them to different datasets and evaluate how it affects the outcome of your model. In doing this, you get to learn what strategies are effective for different datasets and essentially eliminate things that don’t work.
Here’s my article on Feature Engineering to get you started…
Characteristics: This stage requires plenty of conscious effort. In fact, it may feel extremely awkward to begin with and require lots of little adjustments to improve your performance. Completing task may take you hours, days or a week longer than experienced people, but it is key to remember that once upon a time they were also like you and they went through this phase.
Goal: The goal here is to put together lots of small skills that in turn will accumulate to make you unrecognizable to your peers in the future.
Autonomous Stage
Woooosaahhh… The state of Flow.

We hear about it in the podcast, The TED talks, etc. The state of flow what we all crave.
In this state we can perform at maximum levels of proficiency. We aren’t so focused on the skill because we can do it automatically – meaning we don’t have to think about it, it just flows.
Well where does this fit in, in Data Science? Think of it like this, have you ever seen when a new competition starts on Kaggle and it’s always the same people that shoot straight to the top of the leader board? Yes, them! They are in flow. They have learnt about competing on Kaggle, what works and what does not. They have practiced, over and over on countless competitions and now what you see when they shoot to the top is the fruits of their labor.
The skill has become programmed into your subconscious mind and you can now direct your focus to other aspects of your performance.
Conclusion
This process can take a long-time but trusting the process makes the journey lighter. Although a key thing to remember is that the automatic phase will reinforce any bad habits that are picked up along the way and bad habits are very hard to change. Learning best practices from seasoned professionals is always the best way to start rather than having to come back to it.
If there is anything that you think I have missed or some points you don’t agree with, your feedback is valuable. Send a response! If you’d like to get in contact with me, I am most active on LinkedIn and i’d love to connect with you also.
Here are some of my other articles that you may find interesting…