Fuelled by big data and AI, demand for data analytics and data science skills is growing exponentially, according to job sites.
As companies are searching for approaches to harness the power of Big Data, technology professionals who are experts in data analytics and data science are in high demand. The supply of skilled applicants, however, is growing at a slower pace which makes these sorts of jobs great for career changers.
In this blog post, I will assume that you are not straight out of high school, but either already have a university degree (not in computer science or statistics) and/or have already worked for several years and are now considering how to become a qualified data analyst or scientist. There are, of course, many factors to consider when deciding how to enhance your current skillset like previous experience, financial resources, and how much time you want to invest into this, and my intention is to outline the pros and cons of the most common ways to upskill: a university degree, a crash-course (like a Summer School), and online courses of various types.
I am currently a teaching fellow at University College London in the Computer Science department, have taught a Summer School in Data Science in the past, and did a lot of self-learning because my Bachelor’s degree was in Law, so I believe I can give a good overview of the pros and cons on all of the above.
Master’s degree in data science, machine learning, or data analytics
Pros:
If you have the necessary quantitative background and the financial resources (or access to a student loan) then doing a Master’s in data science, machine learning, data analytics, or something more specialised like financial computing, definitely has its benefits: you will have access to a variety of modules as part of a structured course, meet leading academics in the field, meet interesting people and make new friends and (hopefully, in a post-COVID-19 world) experience student life.
Cons:
The main drawback of university degrees (in the UK) is, obviously, the cost. If you are considering studying full time, you do not just need to cover the tuition fees but also living costs for a whole year. A second problem is a fact that it is an application based and you might not be accepted for the programme that you are interested in. In order to do a Master’s degree in data science or machine learning, you need to have a Bachelor’s degree in a quantitative subject. So unless you have studied maths, engineering, economics, or finance, you will not be eligible for such a Master’s programme and will have to do a conversion course first (MSc in Computer Science). For someone, who has no background in computer science, but has the financial resources and the time, I can highly recommend doing a conversion course, as I did back in 2015, after finishing my Bachelor’s degree in Law. But this is, for now, out of the scope of this article.
Another con is that by the end of the degree you might not necessarily have a portfolio of different projects which you could show to potential employers. The coursework often does not reflect real business problems and your Master thesis might be more on the academic, “researchy” side.
Summer School/Intensive short course in data science/machine analytics
Pros:
For people with little or no previous experience in data analytics and data science, a crash course (2-4 weeks) provides the perfect opportunity to learn Python or R, brush up on statistics and hypothesis testing, learn how to work with (numeric) data, and learn the basic concepts of machine learning. Courses like this can help to understand whether you really want to pursue this career path and can help to decide what next steps to take (e.g. a university degree or another course that goes into more depth).
The cost and time commitment of such courses are low and hence are accessible to everyone.
Cons:
Such short courses are only introductory and do not provide you with all the knowledge and tools you need in order to be able to work as a data scientist or data analyst. Many of the concepts are covered in a black-box manner, without going deep into the maths. You might learn how to pre-process your data and feed it into a Logistic Regression classifier from scikit-learn, but not how Logistic Regression actually works.
Self-learning
By self-learning, I mean a structured online course you can do at your own pace, which takes several months to complete, includes assessments and, preferably, has some sort of mentoring scheme.
Cons:
It can be quite hard to find a suitable course which is worth the money. During my research for this article, I have discovered that for many courses (which are not cheap!) the syllabus is not online and needs to be requested, after which you get spammed.
Many courses that are being advertised as “Data Science” courses are actually courses on Data Analytics because they do not cover any maths, but rather databases (SQL) and data visualisation methods. So, if you are pursuing a career in Data Analytics, those courses will cover everything you need. Whether they cover everything a Data Scientist should know, is questionable. I did not find a single course (in English) that covers Linear Algebra and Calculus. Interestingly, when I searched for a course that does cover the necessary maths, the only one I could find is taught in Russian – the same language as the courses on Data Science and Data Analytics where the syllabuses appeal the most to me and which I have recommended to friends and family in the past (who are all very satisfied with them).
Given that you can complete these courses at your own pace, you need to have the necessary time management skills and motivation to finish them, since you often do not have hard deadlines or exams like you do at university
Pros:
The fact that you can do these courses at your own pace can, of course, also be a pro. You do not have to leave your job and do them in the evening or on weekends.
Many of these courses are very reasonably priced and hence accessible to everyone (I would stay away from courses that are priced in the 5-digit range).
In the end, you are more likely to have a portfolio of industry-based projects that you can show to employers. Some of those programs even work closely with industry partners and might connect you to potential employers.
Below I list some online courses where I found the syllabus appealing and the pricing reasonable:
IBM Data Science Professional Certificate (Note: More Data Analytics than Data Science)
Machine Learning with Python: from Linear Models to Deep Learning (Note: full course here)
