Featured Articles

Level Up 7 Data Science Skills with These YouTube Channels

What If Learning Data Science is a Game

We are all familiar with the modern game design that champions or heroes are always equipped with certain attributes and specialties. For example, Dota heroes are scored on the aspects of agility, intelligence, and strength. To excel in the battlefield, the hero needs to have above-average scores among all attributes while additionally specialized in at least one.

So what if we think of learning data science as playing a game where all of us possess multi-dimensional abilities. Playing this game demands constantly sharpen our skillset with weapons, trainings or magic potions which resembles learning through reading, tutorials, and of course those YouTube learning resources mentioned later on.

First of all, let's walk through these seven essential skills that guarantee your success in the data science field.

1. Machine Learning Algorithms Understand the underlying theory behind supervised, unsupervised, and reinforcement learning algorithms, such as:

  • linear regression

  • neural network

  • decision tree

  • KNN

  • clustering

2. Statistics & Math Stats and math are the building blocks of data science, especially in machine learning and AI, including fundamental knowledge of:

  • linear algebra

  • calculus

  • probability distribution

  • hypothesis testing: t-test, ANVOA, correlation

3. SQL SQL is the language used to communicate with the database and derive insights through data extracts and queries, some essentials techniques include:

  • CRUD - create, read, update, delete

  • filter, sort, aggregate

  • date, string, number manipulation

  • join and union

  • subquery

If you would like to get into more details, these articles may help:

Learn SQL in Everyday Language

Get Started with SQL Joins

4. Programming There are some easy to start yet powerful programming languages such as Python and R. Instead of focusing on the coding syntax, more importantly, is to learn the programming logics as well as the developer mindset:

  • loop structure: for loop, while loop

  • conditional structure: if ... else statement

  • data structure and complexity

  • object-oriented programming

5. Data Visualization Data Visualization is embedded throughout the data science journey, from the exploratory data analysis in the beginning to the final reporting and deliverables. Some commonly used tools are:

  • Tableau

  • PowerBI

  • seaborn (Python package)

  • ggplot2 (R package)

6. Project Implementation Understanding the theory and concepts is crucial, but implementation is also imperative. This skill is more focused on how to put knowledge into practice by building projects and implement the data science lifecycle:

  • Business Understanding

  • Data Mining

  • Data Cleaning

  • EDA

  • Feature Engineering

  • Predictive Modelling

  • Data Visualization

7. Storytelling This is the soft skill that is often neglected. This determines whether the data scientists are capable of conveying the message succinctly to their audiences while also keeping them engaged. Since data science is building the bridge between business and technology, being able to articulate complex techniques and processes to people from various disciplines becomes an essential skill to have.

What YouTube Channels to Follow?

For beginners, there are many channels out there providing tutorial contents that are extremely helpful, but to some extent, make data science seem daunting. Ultimately, the learning journey becomes more like a chore. To gamify the learning process, we need those YouTubers who embed their own personality and creativity into their channels. I really appreciate them putting the effort to make the contents more accessible and engaging for their audience.

In the Data Science game, if we want to level up the skillsets, it requires consuming the appropriate resources accordingly. If we continue with the Dota analogy, there are weapons such as swords mainly increase your strength whereas boots enhance your agility. Similarly, these channels each has its own unfair advantage. Combining them together you will be able to hone your skills in a holistic way.

1. Ken Jee

His channel is very project-focused and beginner-friendly. It's a great place to get started with building data science projects, especially Kaggle projects, and not intimidated by the math or statistics behind the complex algorithms. Ken Jee also provides useful career tips and productivity hacks.

2. Joma Tech

This might be the most random but creative data scientist I ever followed. It does have the magic that makes you keep watching his videos. Joma Tech describes data science from a programmer's perspective. For example, he has a series called “If Programming Was An Anime” which reaches millions of views. His vlog-styled contents will surely let you be entertained and educated at the same time.

3. StatQuest with Josh Starmer

This channel focus on illustrating machine learning concepts and algorithms through animated visuals. It is amazing how he breaks down complex concepts (e.g. Stochastic Gradient Descent, Support Vector Machine) into digestible pieces. It is my go-to channel whenever I need to learn a new ML model or get stuck with a complex algorithm.

4. 3Blue1Brown

This channel is an excellent combination of arts and science. The creator Grant Sanderson narrates the story of the mathematic world through stunning visual illustrations from a unique perspective. There is a series about the math and probability behind COVID-19 which I highly recommended.

5. Nate at StrataScratch

It provides thorough walkthroughs of SQL interview questions from big tech companies such as Microsoft, Facebook etc. For those that are preparing for data science technical interviews, you may want to check it out. The exercises help to consolidate the SQL implementation through the process of active recall. If you would like to know more about how active recall may assist in learning, check out my article.

6. Art of Visualization

What makes this channel stands out is that it covers a range of custom charts that are not intuitive to create in Tableau, including Sankey diagram, Sunburst chart. Additionally, there are series of tutorials surrounding the topic of data visualization using Python, R, and beyond.

Take-Home Message

This article covers seven essentials skills for data scientists and also recommends six YouTube channels that assist in learning these skills. Hopefully, you may find these resources helpful for your learning journey. As always, let's continue to level up our skillsets together.

249 views0 comments

Recent Posts

See All