Preparing for a Data Science/Machine Learning program
If you are reading this article, there is a good chance you are considering taking a Machine Learning(ML) or Data Science(DS) program soon and do not know where to start. Though it has a steep learning curve, I would highly recommend and encourage you to take this step. Machine Learning is fascinating and offers tremendous predictive power. If ML researcher continues with the innovations that are happening today, ML is going to be an integral part of every business domain soon.
Many a time, I hear, "Where do I begin?". Watching videos or reading articles is not enough to acquire hands-on experience, and people become quickly overwhelmed with many mathematical/statistical concepts and python libraries. When I started my first Machine Learning program, I was in the same boat. I used to Google for every unknown term and add "for dummies" at the end :-). Over time, I realized that my learning process would have been significantly smoother had I spent 2 to 3 months on the prerequisites (7 to 10 hours a week) for these programs. My goal in this post is to share my experience and the resources I have consulted to complete these programs.
One question you may have is whether you will be ready to work in the ML domain after program completion. In my opinion, it depends on the number of years of experience that you have. If you are in school, just graduated or have a couple of years of experience, you will likely find an internship or entry-level position in the ML domain. For others with more experience, the best approach will be to implement the projects from your boot camp at your current workplace on your own and then take on new projects in a couple of years. I also highly recommend participating in Kagglecompetitionsand relateddiscussions. It goes without saying that one needs to stay updated with recent advancements in ML, as the area is continuously evolving. For example, automated feature engineering is growing traction and will significantly simplify a Data Scientist's work in this area.
This list of boot camp prerequisite resources is thorough and hence, long :-). My intention is NOT to overwhelm or discourage you but to prepare you for an ML boot camp. You may already be familiar with some of the areas and can skip those sections. On the other hand, if you are in high school, I would recommend completing high school algebra and calculus before moving forward with these resources.
As you may already know, Machine Learning (or Data Science) is a multidisciplinary study. The study involves an introductory college-level understanding of Statistics, Calculus, Linear Algebra, Object-Oriented Programming(OOP) basics, SQL and Python, and viable domain knowledge. Domain knowledge comes with working in a specific industry and can be improved consciously over time. For the rest, here are the books and online resources I have found useful along with the estimated time it took me to cover each of these areas.
Before I begin with the list, a single piece of advice that most find useful for these boot camps is to avoid going down the rabbit hole. First, learnhow,without fully knowingwhy. This may be counter-intuitive, but it will help you learn all the bits and pieces that work together in Machine Learning. Try to stay within the estimated hours(maybe 25% more), I have suggested. Once you have a good handle on how you will be in a better position to deep five into each of the areas that make ML possible.
Machine Learning:
SQL:
OOP:
Python:
Probability & Statistics:
Calculus:
Linear Algebra:
These are the math and programming basics that are needed to get started with Machine Learning. You may not understand everything at this point ( and that is ok), but some degree of familiarity and having an additional resource handy will make the learning process enjoyable. This is an exciting path and I hope sharing my experience with you helps in your next step. If you have further questions, feel free to email me or comment here!