The best way to get your feet wet is to do a real project that meets a real requirement. If you ask around, there are plenty of people, local business, mom-pop places that have some sort of pain point that can be solved with a tool. This is how you get the idea for the project.
Get really focused and deep into making this tool. Use the appropriate resources for the tool, don't be completely brainless (if the tool is a site, obviously use a language that produces markup for example). Don't worry about "broadening" your learning and just REALLY focus on making this tool address the problem then once you are done, give it to them and ask for feedback. Based on that feedback, refine it then give it to them again.
Once you got an understanding of how to go from a problem to a solution, then you can start broadening your understanding of coding including the development process, requirements, design, implementation, testing, validation, apply it to what you did and how you could have done it differently/better. Because you went deep into your tool, now is a good time to broaden your understanding of the concepts you needed for your tool (structures, complex conditions, algorithms, whatever it is).
The point is, my philosophy of "jumping" into the coding world is from a "bottom-up" approach. You focus and go deep on a very specific a limited and narrow task then slowly add the layers of understanding from there and move your way up to the bigger and broader picture. Too many times, I see people take the high level approach and learn about concepts, patterns, methodologies, project management, theory, etc, all before they do any practical implementation and end up being overwhelmed or simply just not "getting it" and go into a career of being an analyst instead.
TL;DR: Don't be an analyst.