Continuous Learning: A Data Scientist's Odyssey

To be a data scientist is to sign up as a lifetime learner. Something new is always coming up in the data science field – – a new algorithm, a new practice, a new concept. How should we data scientists keep current and navigate through this ever-changing field? In this article, I would like to share what to learn and how to learn continuously as someone who is self-taught in data science and currently works in this field.

Why learning continuously

Before discussing how to learn as a data scientist continuously, it’s important to first understand why. Learning without purpose seldom gives an actual uplift in skills since it’s hard to keep you motivated consistently. For me, continuous learning is almost a necessity at work. Encountering roadblocks at work is common due to the complex nature of the problems. Building something from scratch or improving problem-solving methods always requires extra research, reading, and practice. Besides, I am also a curious person and eager to learn more about the latest trends and cutting-edge technologies in my field. Going through my past projects and articles at Medium, some techniques that were cutting-edge back then look so outdated today. I am amazed by how fast this field is evolving, and at the same time, I feel obligated to evolve and update my skill sets. For me, continuous learning increases work efficiency, boosts confidence and job security, and inspires me in content creation. In fact, a mindset of continuous learning is beneficial for anyone seeking growth and willing to step out of their comfort zone, not just limited to data science or the tech industry. Find your ‘why’ before you start.

What do I learn

Deciding what to learn is important. It depends on your current skill level and your short-term and long-term goals. I focus on learning two types of skills: hard skills to generate insights and soft skills for delivering those insights effectively.

Hard skills

Data science relies heavily on hard skills, which are the fundamental abilities necessary to excel in this field. It’s essential to have solid mathematical and theoretical foundations and stay abreast of the latest algorithms and best practices in coding. From a practical perspective, knowing the math behind an algorithm or a technical concept seems to take a lot of time and effort, but I still find it beneficial in the long term. On top of these, problem-solving is another important skill to have, which involves quickly learning, navigating ambiguity, summarizing, and generalizing specific tasks for future use. These skills will equip you to generate useful insights and business value from messiness and ambiguity.

Besides these general skills, domain knowledge is also very important as you gain more experience in specific industries. Businesses operate very differently across industries. Thus, the way of driving insights can be very different. For example, data analysis in healthcare is very different due to the privacy concerns of the data and how data are collected and stored. Industries like E-commerce or social media may have a lot of data and relatively easier to run experiments to get insights, while traditional industries might have more limitations. It’s important to connect data science with domain knowledge in your industry by understanding how the business operates.

Soft skills

Soft skills are as important as hard skills in career and personal growth. They are the skills that distinguish you from other data scientists and help you to go further. The soft skills I keep improving include effective communication, storytelling, presentation, leadership, business acumen, etc. No one is an isolated island in the workforce. It’s not enough to just generate insights. You must also learn to effectively communicate and collaborate with others to maximize impact. The following article discusses more details in these skills:

Seven Principles I Follow To Be a Better Data Scientist

Where do I learn

At work

Learning by doing is the most efficient way of learning and enhancing a skill, and your coworkers are the best resources. At work, you are learning new skills every day (hopefully). You are practicing hard skills by solving problems, delivering results, and developing soft skills by communicating and collaborating. To learn at work, maintain an open and positive mindset, set goals that promote growth and consistently reflect on what you have learned and practiced recently. This will demonstrate that the years of experience listed on your resume are more than just numbers, but a reflection of the progress you’ve made in developing your skills.

Medium

Medium is probably my best investment, with the highest ROI. I read blog posts to have a good understanding of new concepts and keep up to date with the latest practices in the industry. It is much faster to grasp concepts than reading a paper or taking a course. I recommend following technical blogs and subscribing to Towards Data Science, Analytics Vidhya, Towards AI, CodeX, etc. Besides, tech companies’s tech blogs are gems. I follow Microsoft, Netflix, Airbnb, DoorDash, Expedia, Pinterest, etc. These are the ones update frequently.

I also use Medium to network. I get to build a great community with my articles and get connected with other talented writers. My readers can see my progress over the years based on the content of my articles, from just entering the field of data science, to working on complete projects, to looking for a job in the field, and now on sharing experience and learnings from work. Connect with me if you want to read more articles about data science techniques and career development.

Online learning platforms

If I want to learn something new in-depth, I will sign up for a course on online learning platforms like Coursera, Udemy, LinkedIn learning, etc. Even though watching YouTube videos is free, sometimes they are not enough for me. If I sign up for a course, I want to learn thoroughly and have a lab to practice and receive feedback. As mentioned in my previous article, deliberate practice is crucial to learning new skills. Practicing and getting evaluated at labs ensures that I understand new concepts well and am prepared to apply them when needed. Companies usually provide employees with a learning budget for these online platforms, so make sure to use them to invest in yourself. Plus, if you are also based in the US, you will be amazed by how many free resources your local library provides. I get free Coursera, LinkedIn Learning, and SkillShare accounts from my local library. Gather what’s offered in your local area’s library website.

Paper and Books

Reading research papers or textbooks seems inefficient and overwhelming. However, it is one of the few ways that help you learn something thoroughly and structurally. For skills you want to specialize in or fields you have a deep interest in, having trustworthy textbooks as references and keeping up with the latest research papers are great choices. The only inefficiency comes from how you read these materials. It is a waste of time just to read them sentence by sentence. There are some advice I want to share to read with efficiency:

Read with questions. Why do you open this book, or why do you need to read this paper? If you have particular questions in mind, your reading experience will be much more quick and direct to the point.
Read the abstract and introductions first for research papers. Research papers can be heavily in math and details that might not be relevant to you. The abstract and introduction should be sufficient if you just want to explore what’s new. You can always read more thoroughly if you are interested.
Find slides or syllabi for books or papers. There are often slides available summarizing the key points of famous books and papers.It is much easier to read through a slide than the whole body of text.
Find YouTube videos or abstracts for textbooks. Some textbooks have supplementary videos online either provided by the book author or by professors teaching a class with this book. Videos can be an alternative medium that helps learners absorb knowledge in a visual way, which some might prefer. You can also get the abstract and summary of a book either by searching online or through getAbstract:

getAbstract: Know better. Do better.

Ask AI to summarize them for you. You can also use tools like ChatGPT to help provide key points of a PDF format file if you only want to know the high-level summaries.

Conference and Webinars

Attending conference or webinars are an efficient way to stay updated with the latest technology development and practice. It is also a good chance to network and connect with your community, especially for the in-person ones. At conferences, attendees are exposed to the latest developments in various areas related to a specific topic at a high level. For topics of personal interest, speakers often share their source code and papers online, providing a great opportunity to connect and collaborate. Attending these events has always been an eye-opening experience for me. I also feel incredibly motivated after interacting with so many brilliant minds. There are a lot of conference choices for data scientists. Check the theme and speakers to go to the one that benefits you the most now or in the future. You can choose based on locations, communities, or specific topics. Also, it would be great if your company could cover your expenses. Convince them to invest in you.

Podcast

I follow a lot of great podcasts talking about the latest trends and practices in data science and the industry. I also like these podcasts that give insightful advice on improving productivity, promoting continuous learning, career development, etc. Here is a list of podcasts I would recommend:

Data Skeptic by Kyle Polich

The Data Skeptic Podcast features interviews and discussion of topics related to data science, statistics, machine learning, artificial intelligence and the like, all from the perspective of applying critical thinking and the scientific method to evaluate the veracity of claims and efficacy of approaches.

Super Data Science: ML&AI Podcast with Jon Krohn

We cover tools, techniques, and implementation tricks across data collection, databases, analytics, predictive modeling, visualization, software engineering, real-world applications, commercialization, and entrepreneurship − everything you need to crush it with data science.

The Data Scientist Show by Daliana Liu

A deep dive into data scientists’ day-to-day work, tools and models they use, how they tackle problems, and their career journeys. This podcast helps you grow a successful career in data science. Listening to an episode is like having lunch with an experienced mentor. Guests are data science practitioners from various industries, AI researchers, economists, and CTOs of AI companies.

A Bouquet of Arguments by Y&M

For those who understand mandarin, I highly recommend this podcast for career advice, especially in tech. They discuss a lot of career myths and how to improve soft skills at work. Plus, it’s gym-friendly!

Huberman Lab by Dr. Andrew Huberman

Huberman Lab discusses neuroscience: how our brain and its connections with the organs of our body control our perceptions, our behaviors, and our health.

Huberman Lab is more like a productivity podcast for me. I like how Dr. Huberman’s evidence-based suggestions. The various topics expand my knowledge horizons.

Lex Fridman Podcast by Lex Fridman

Conversations about science, technology, history, philosophy and the nature of intelligence, consciousness, love, and power.

Other Podcast that discuss specific industries’ development and news

As you progress in your career, domain knowledge becomes increasingly important in addition to general skills. If you are in a specific industry or interested in some, follow some of the podcasts to keep updated on the latest trends and practices.

Podcasts provide an alternative medium to deliver knowledge and information. You can listen to them during the commute, at workouts, or doing house chores. While some can be concise, others can provide technical details and information with great density. I must admit I should put more time into listening to podcasts by finding a better time. I usually listen to podcasts at the gym doing weight training or running, and it’s probably not the best time to get super technical. If you find a good time to listen to informative podcasts, please do share with me!

In summary, there are a lot of ways to learn. We can learn at work and outside of work, the ways you learn barely matters, the most important thing is to keep a growth mindset. Take an extra step to summarize and generalize so you can internalize them into your skill sets.

What are the actions?

Having a growth mindset is important, but taking action is equally crucial. Here are some tips to maintain a habit of continuously learning:

Set up long-term and short-term goals

Goals give us the why of our actions, which motivate us when we lack willpower. In time management theory, it is recommended to categorize tasks based on their importance and urgency. While urgent tasks are always prioritized, delegating time for important but not urgent tasks every day is essential since that sets you apart from others in the long run. These tasks include learning a new language or skill, starting a side project, etc. What you defined as these tasks will depend on your long-term and short-term goals. These goals should not be limited to work or school. You can also include goals in personal growth, relationships, etc. Reflect and ask yourself:

Who would you want to be in the next year and in the next ten years?
What makes you happy and satisfied?
What excites you the most?
What do you value the most in life?

These questions will help you find your goals.

Set up fixed times to learn

When I was at graduate school, I was used to taking classes, working on research projects, and teaching at the same time. I didn’t enjoy a lot of full weekends without feeling guilty. I didn’t count how many hours I work weekly, but it’s definitely over 40 hours. Once I started working, I realized that working side hustles or learning something new requires more effort when juggling a full-time job. Therefore, I have learned the importance of scheduling specific time blocks on my calendar outside of work hours to focus on these tasks. Otherwise, I may find myself working hard one day and don’t what to do the next day. Or mindlessly wasting time on the couch without even realizing it. I usually block several weekday evenings for working on side projects or taking online courses and mornings for reading interesting blog posts. Find the schedule that works the best for you to try it out, and then make adjustments accordingly until you find your comfortable weekly routine.

Do not use Read Later but never read later

Many online platforms have a function that allows users to collect something and read it later. It is convenient if you don’t have time to finish reading or want to read deeper later. However, let’s be honest: how often do you revisit the collections? How many times have you taken a long time to find a paper, a course, or a textbook online, but once you found it, you just let it sit in your local drive but never really read it? How many times have you added a Medium story to your reading list and never opened it anymore? I catch myself almost never reading the articles in my reading list. It is understandable that nowadays, with easy access to social media, we can be overwhelmed by the information we receive every day. We need to keep an open mind for new information and knowledge and know what we should prioritize receiving. The collect and read later function helps filter the information we are interested in, but we need follow-up steps to grasp and absorb. I found it useful to dedicate specific times of the week or day to reading things in the collection.

Additionally, categorizing your collections is super useful. Instead of putting everything in my "Reading List," I put articles by topics like "Causal Inference" and "Generative AI" or by projects as supporting resources, like "Time Series class resources." I found myself revisiting specific topics more frequently for specific purposes. Little organization makes a big difference.

Share what you have learned

Teaching or sharing what you have learned is a great way to ensure you have learned something correctly and thoroughly. Like how I started writing at Medium, initially, I wanted to keep the blog posts as references and notes for my interview preparation. While writing down what I have learned, I am forced to really think these concepts through without any ambiguities. When we are reading or attending lectures, we take in or input new information. It is essential to take out or output the information for us to close the learning loop. Writing blog posts not only reinforces my learning but also inspires others. I highly recommend starting the outputting process, no matter in the format of writing, video, discussing with learning partners, or teaching a grandparent.

Method vs Methodology

One useful insight I learned from listening to the Podcast A Bouquet of Arguments is knowing the difference between method and methodology. When completing a task, we follow a method to get this specific task done. In the meantime, we summarize a methodology to get all tasks similar to this one done in the future. For example, you are tasked to clean a specific data set before the EDA and modeling process. You preprocess this dataset by following methods of filling in missing data, checking duplicates, etc. You might need to customize the method due to specific data formats. After several similar tasks, you might get a generalizable method for preprocessing most datasets, like what this article has summarized. Learning through summarizing and reflecting internalizes the task-solving process into your skillset. Without this crucial step, you miss out on the most important aspect of learning by doing, which extends your ability to solve specific tasks to solve a wide range of similar tasks in the future.

Make the best use of available tools

Using the right tool is not being "lazy"; rather, it’s efficiently allocating limited energy and attention. AI is a good study pal, consultant, teacher, assistant, etc. Use it responsibly and critically. The latest technology offers many advantages that we can benefit from. Here are some great medium articles I came across that share the amazing tools that we can all leverage to learn and practice data science better:

How I Would Learn Data Science with ChatGPT (If I Could Start Over)

ChatGPT is Old News: Here are 8 AI Tools that Will Transform Your Work

How to Use ChatGPT to Learn Data Science Faster, Even If You Are Already Advanced

I will also discuss more aspects of using AI in data science self-education in future articles. Please stay tuned for further updates.

Watch motivation videos

I follow many YouTubers who produce videos about productivity, time management, or study vlogs. Watching their videos motivates me to get together when I feel lazy and unproductive. You can also search for recorded x-hour study-with-me videos from different YouTubers. They always have a nice setup with good music and good vibes. Sometimes, I open one as my Pomodoro Timer to help me get focused when doing deep work. There are also study/work live streams that you can join. It will make you feel like in a classroom or library with good communities that hold each other accountable.

Give yourself breaks

Maintaining a full-time position and continuous learning outside of working hours sounds scary. Sometimes, it is indeed overwhelming! Burnout is so common in the tech industry that you are entitled to give yourself breaks occasionally. Know when to say no and when to give yourself a break. You can have vacations and plan a resting day of the week that you use to do activities that relax you the most. I relax by going to the gym or visiting a bookstore, and I take PTOs every quarter to travel somewhere. Knowing how to relax is also a critical skill as important as knowing how to be productive.

One may ask, maintaining a full-time position is already tiring, do we really need to push ourselves too hard? Recently I read a book called The 4-Hour Workweek by Tim Ferriss. With all the refreshing points brought up by the book, one I want to share in this article is "avoid distress but embrace eustress." Distress is bad, it makes people burn out, however, we all need "eustress" in life. Eustress is the healthy kind of stress, the one that gives you energy and motivates you to be better. Continuous learning provides me the healthy stress, the one that doesn’t give me anxiety but fulfillment. Thank you for reading this far. Hope this article is useful for you to find your comfortable ways of navigating in this ever-change field.

Disclaimer: All recommendations are based on my personal experience. There are no ads in this article, but I welcome future brand sponsorships and collaborations.

Lastly, don’t forget to: