PhD to Data Science? Avoid These Mistakes.
With all the blues going on in academia (lack of funding, mental health risk, and scarcity of gainful academic careers), and all the buzz around Data Science many PhD's consider Data Science as an alternative career path. If you want to avoid some of the common pitfalls spend 5 minutes and read this.
You can listen to my interview with Cheeky Scientist Radio on what was my journey to Data Science.
DON'T DO IT IF... Or Is Data Science For Me?!
Don't like mathematics or don't have good foundation in linear algebra and multivariate calculus? Then perhaps this is not a space that you want to be in. While coding and playing around in Python, R, Scala, and ... is very fun (if you don't find it to be fun, then you have your answer and stop reading the rest of this post), at the end of the day you have to read a lot of papers that are heavy on mathematics.
Don't like staring at a monitor and code or clean data? This will be a relatively big part of your day. If you don't enjoy it you won't thrive in this field.
Don't like talking to "laymen"? You have to communicate your results in a way that is actionable for the business. Very few people may care about how you got the result but you have to be able to communicate it very clearly. If you don't like talking and working with people with different backgrounds, this is not a great career for you.
DON'T Chase the Technologies, Get the Foundations Right!
Rather than chasing different technologies try to get the basics solid. R, Python, Scala, Hadoop, Spark, Cloud Computing, and ... they are all good skills to have but don't keep chasing them. Instead, first ensure that you know the basics of your stats and algorithms well before trying to answer is it better to learn R or Python.
Online Courses are great but NOT Enough:
Online MOOC websites have revolutionized learning by democratizing it. They are very good to get you started but by no means enough to prove to your future employers that you can help their business solve their problems. You need to expose yourself to real world problems were the data is not well curated, the question is not clear, and you may have to even define the success criteria! One way of improving your skills and get a better feel of what you will face in business is through consultation jobs/gigs. Start your own consultation gig and see what are the problems that a business owner have. Try to get your hands dirty by looking into those problems.
DO Interview As Much As You Can:
One way to hone your interviewing skills and more importantly finding the gaps in your knowledge is through interviews. When you do a job interview, you have the opportunity to sit with an expert for half an hour to several hours and learn from them. It helps you find the gaps in your knowledge.
Basically, you study and practice new skills, you progress, you interview, you use the feedback, you study more, and you become better and after some iterations you become a better data scientist.