Should I Learn Data Science by Doing or First Study the Theory?

Learning by doing definitely helps things stick better! Make sure to utilize debugging tools; they can deepen your understanding significantly. If you're working on algorithms, try jotting steps down on paper or modeling them in Excel to see how values change.

I prefer to learn by doing after some foundational understanding. This way, I have a grasp of what I'm doing, know what results I should expect, and can identify where I might make mistakes. Learning without a clear understanding can feel inefficient. You might want to try both methods and see what works best for you!

In a traditional learning environment, you often get stuck in a cycle of passive training—just sitting and listening. The goal should be finding a balance that encourages more hands-on experience than you might expect. The more you actively apply what you learn, the better it will stick with you.

I think a mixture of both approaches works well. Start doing what you can, figure out what gaps you have in your knowledge, then learn those concepts, and repeat the cycle. For data science specifically, getting a solid base in linear algebra, calculus, and statistics is crucial before diving into practical applications.

Your learning approach should match your objective. If your main goal is personal growth, jumping in and learning through mistakes can be very educational, albeit risky. But if you're aiming to accomplish something specific, it might be wiser to invest time in understanding the theory before executing projects.