3 Practices for Growing your Data Science Career, Building Expertise, and Earning Reputation

Photo by Saulo Mohana on Unsplash

“How can I be successful in my data science career?”

I often asked this question to myself and several top data scientists in the Industry. In the past 5 years, I had the opportunities to lead multiple data science projects at IBM, my workplace. I worked closely with many fine data science professionals, including data scientists, machine learning engineers, data engineers, and data science leaders. I mentored junior data scientists and was mentored by veterans.

From reflecting on my career and advice of the data science colleagues, I have recognized three practices that can help data science professionals refine their art and grow reputation as experts in their field.

Practice 1: Learning from the best

Data Science has a massive volume of learning resources, which include online courses, books, blogs, and code examples. Not all of these are of outstanding quality. Since we have limited hours in a day that we can spend for learning, it’s important to find and use the best learning resources for an optimal learning outcome.

Online Courses

Andrew Ng, a Professor at Stanford and a co-founder of Coursera, is one of the most popular instructors for teaching machine learning. I have taken three machine learning courses with Andrew Ng, via Coursera. He clearly explains the core concepts. Before watching his lectures, I would often write down a list of questions I had on the lecture topic. Interestingly, he answers almost all the questions during the lectures.

Andrew Ng’s “Machine Learning”, a Coursera course, helped me grasp a solid foundation of machine learning concepts. As of now, this course has 3.8M enrollments and a 4.9/5.0 rating computed from 153,955 ratings. If you’ve not taken this course, I strongly recommend you complete this course soon. Though it’s possible to audit it for free, I would advise you to go for the certificate option, which cost me $79 in 2017. Paying for the certificate gave me an extra motivation to complete this course. After the completion, I received a nice certificate, which I added to my LinkedIn profile. I completed this course in 50 hours over a 10-week period.

Kaggle

Kaggle — the premier data science competition platform — is a great place to build data science expertise by solving data science problems and competing with others. Among 5 million data science practitioners or enthusiasts registered at Kaggle, the top ones are Kaggle Grandmasters, a title they earned after consistently proving their skills in data science. This link will show you the list of Kaggle Grandmasters. By clicking any Grandmaster’s name, you can go to their Kaggle page and find the code they had shared. Looking at the data science solutions of the top data scientists at Kaggle is a great way to refine your data science skills.

Books

I checked out many machine learning books and found the following 2 most useful:

1. Python Machine Learning, 3rd Edition, by Sebastian Raschka and Vahid Mirjalili
   This book does an excellent job in explaining the core machine learning concepts and providing plenty of high-quality code examples. Its accompanying Git repository has an excellent collection of working code examples on a wide range of machine learning topics.
2. An Introduction to Statistical Learning, by James, Witten, Hastie, and Tibshirani
   The book provides a clear and concise explanation of the many essential concepts in machine learning. You can download this book for free from here.

Practice 2: Following Latest Research

For staying current, it’s important for data science professionals to get into the daily habit of reading research publications from major machine learning conferences and journals, and a list of major machine learning conferences can be found here. The format and the theoretical content of the research papers can be overwhelming. During my master’s research, I didn’t know how to approach a research paper. I tried to read each paper from its beginning to the end and often struggled to finish the reading.

I improved my skills in reading research papers by following the strategies from this two-page article How to Read a Paper, written by S. Keshav, a professor of computer science at the University of Cambridge. This article advises not to read a paper from the beginning to the end in one go, rather it suggests taking three passes through a research paper. The following is my adaption of the three passes:

1. Pre-reading: in this pass, I try to grasp a gist of the paper by reading its title, abstract, introduction, section and sub-section headings, conclusion. After this pass, I decide whether I want to take a second pass through this paper. If the paper is irrelevant, uninteresting, or has incorrect assumptions, I put this paper aside and look for another paper. This pass takes roughly 15–20 minutes.
2. Superficial Reading: In the second pass, I try to grasp as much as I can without getting stuck. I do not stop at what I can’t understand; I make a note of it and move on. I see the organization of the paper, skim through the paper and pause when I see something that looks relevant and interesting. I carefully see figures, diagrams, and plots. I write questions and key points. I skip any complex theoretical content, including mathematical proofs. This pass gives me a decent understanding of the research approach and experiment results. With most papers, this level of reading is sufficient for me. This pass takes me roughly 1.5 hours.
3. Analytical Reading: From the previous 2 passes, if I come across an outstanding paper that is very relevant to my research interests, I go for the last pass with the goal of gaining a deeper understanding of it. During this pass, I immerse myself in the paper. I draw pictures and create mind maps as I read the paper. If I see some new concepts that aren’t clearly explained in the paper, I google for them. For me, a beginner researcher, this phase takes several hours to a few days, depending upon the complexity of the paper.

Practice 3: Building your reputation

As a data science practitioner, hard work alone is not sufficient. You need to work consistently towards building your reputation as an expert in your field–both inside and outside your organization. Two brilliant methods of building reputation are writing and speaking.

Writing

Build a regular practice of writing and publishing on data science. Medium and LinkedIn are two great places to post and amplify your messages. For those who need polishing their writing skills, a great writing reference is:

The Elements of Style, by William Strunk Jr. and E. B. White

William Strunk Jr., who was a Professor of English at the Cornell University in the early 1900s, wrote this classic writing reference a century ago. E. B. White-William Strunk Jr.’s student and a famed essayist-later revised and expanded this book. I have been using this reference since 2012. As a non-native English speaker, I have benefited from the rules and styles pertinent to the English language taught in this 100-page book.

An effective writing process can dramatically improve a Writer’s productivity. I follow the Publication Coach writing process outlined in the book 8 ½ Steps to Writing Faster, Better, written by Daphne Gray-Grant. Every Tuesday, Daphne Gray-Grant publishes her free Newsletter, Power Writing, with great tips on writing habits.

The most important lesson I learned from Daphne Gray-Grant is to break writing and editing into two separate steps. Following her advice, when I write, I focus on producing a messy first draft as fast as possible, without pausing and making any edits while writing. Once I have the draft, I let some time pass — typically a night or a day — before I revise and improve my draft. Leaving this time between drafting and revising is important to clear the mind and take a fresh look while editing the draft.

Speaking

Speak often and share your knowledge with the world. The more you speak, the better you get at it. In parallel, your reputation as an expert in your field will grow. During 2020, I actively searched for speaking opportunities both inside and outside of my organization. Simply by asking my colleagues for speaking opportunities, last year I spoke at a few global conferences and webinars–over 500 people listened to my talk live, and many more watched the recordings. While preparing my presentation slides, I follow the tips from the following two books:

1. HBR Guide to Persuasive Presentations, by Nancy Duarte
2. Presentation Zen: Simple Ideas on Presentation Design and Delivery, 3rd Edition, by Garr Reynolds

Both are popular books on this subject and have great tips on preparing and delivering a great presentation.

Final Words

Amidst our daily busy schedule, it can seem difficult to find time to think and act on our career. However, if we can spare a certain amount of time every day to progress our career, the payoff over time will be significant. In 2021, I made a commitment to spend at least 30 minutes every day to nurture my data science career. How about you?