How to Get Started on Your Data Science Career Journey

Six considerations for beginners to pick a resource for upskilling in Data Science and AI/ML

Introduction

To state the obvious, data science has evolved into one of the most sought-after skill sets in the market over the past decade. Traditional corporations, technology companies, consulting firms, start-up businesses — you name it — are continually hiring data science professionals. High demand and a relatively short supply of experienced experts in this space make this a very lucrative career opportunity. To break into and be successful in this area, you need a deep understanding of not just the available algorithms and packages but also develop an intuition around which methods lend themselves to which use cases. Plus you’ll need to learn how to translate a real-world problem to a data science framework. In this post, I’ll talk about how beginners can build a fundamental and deep understanding of this space to initiate a career in this area.

Where do you start?

Given the plethora of resources available to learners, it can be confusing as to what to pick as a learning mechanism. While it would depend on your individual situation and objective, you may want to consider a few criteria when selecting a learning resource:

· Content: typically, a robust resource will cover supervised learning (e.g. linear regression, logistic regression, decision trees, ensemble methods), unsupervised learning (e.g. clustering, PCA), as well as basics around statistics and probability. Many programs have focused modules on advanced methods including Deep Learning, Computer Vision, Natural Language Processing, and Generative AI. Some programs even offer free previews of the course contents, learning videos, and coding projects to help learners make a decision. While it may be challenging to find a program that includes every topic in detail, your purpose for pivoting into the data science domain should dictate the choice of the coursework.

· Audience: “Who will benefit from the coursework?” is a key question to help with the determination of a program. If you are a beginner, intermediate level courses may prove to be a barrier to learning. On the flip side, very basic concepts may be less useful for learners who already have some background in analytical methods.

· Time commitment: depending on your individual situation and preferences, this consideration may narrow the options you may have. For instance, if you are a working professional, you may not be able to spend beyond a few hours a week. Many online programs offer flexible timings with an expected 5–7 hours of work per week targeted at part-time learners. Overall time spent in a program also determines how much skill development occurs. While advanced programs at universities (e.g. master’s in data science) may serve to develop deeper expertise over 18–24 months and may require several hours of dedicated effort during a week, shorter courses over 4–6 weeks may not be helpful in developing the necessary skills to kickstart a new career.

· Learner experience: you may find user reviews on the program page or at other locations online. These can be helpful in getting a feel for the quality of course content and teaching, especially if there’s a sizeable number of ratings as well as comments. Further, the reach of the program is indicated by the number of learners that are either currently enrolled and/or those who have completed it. One could argue that verifying the validity of reviews and number of learners may be challenging. Even so, these can serve as triangulation points, given other information, for looking at a program holistically.

· Cost: program fees can be an important factor in selection. Programs can range from being free of charge to running into thousands of dollars.

· Career resources: oftentimes, programs offer support for a new job search and career switch, including interview prep, resume reviews, and networking with domain experts in industry. While this may not be a primary consideration regarding choice of a program, these can help get a foot in the door as you start on the career change journey.

Suggested learning mechanisms

While there are several online resources you can use to gain knowledge of specific topics, I would recommend for beginners to consider a structured pathway that provides a comprehensive view of the data science area. Based on my experience, I would recommend choosing from three types of programs, at least initially (one can always supplement the learning with additional material on individual topics):

· Specialization or Professional Certificate

· MicroMasters/Nanodegrees

· Bootcamps

Typically, these programs spread over several months. I have found these to offer a balance between single/short courses that may not be sufficient to develop the required acumen in data science, and formal education that may take more than a year. EdTech companies regularly offer these programs, oftentimes, in collaboration with highly regarded universities. A side-by-side comparison across the selection criteria can be helpful in deciding on a path forward — Figure 1 shows an example comparison of specializations and professional certificates. The courses listed here are only a small sample of the variety of programs online and are by no means exhaustive. Furthermore, the details presented for each program are from a particular point in time and may change.

While it is recommended to select and start with a single program, sometimes it may take multiple courses to feel comfortable with the subject. In my own case, I started with Andrew Ng’s Machine Learning course. It covered the fundamentals in good detail but as the course had exercises in Matlab/Octave at that time, I decided to do another course from IBM to learn Python along with AI/ML. Based on my personal learning preference, I felt the need for a more formal mechanism of instruction and decided to go with a Post Graduate Program in AI/ML from the University of Texas at Austin in collaboration with Great Learning. This provided me with more structure as I had theory to learn from video modules taking up to 5–7 hours per week, take weekly quizzes, attend an online mentor-led group interactive session for a couple of hours every weekend, and code submissions every few weeks. While the theory and quizzes helped me gain an understanding of the concepts, the hands-on code development was really my key to internalizing the learning. This exercise gave me fundamental insights on how to set up the problem, clean and analyze raw data, determine which algorithm to apply, refine the solution, and explain the results.

To conclude…

It is important to make an informed decision when embarking on a skill-building journey. As a beginner, it may be prudent to consider a more structured learning pathway, and then deep dive into specific areas once you have a fair understanding of the basics. Spending time upfront in researching and benchmarking programs can also help learners understand the key and in-demand topics within the data science realm.

Thanks for reading. Hope you found it useful. Feel free to send me your comments to [email protected]. Let’s connect on LinkedIn

How to Get Started on Your Data Science Career Journey was originally published in Towards Data Science on Medium, where people are continuing the conversation by highlighting and responding to this story.