In data analysis, concentrating on getting extremely good at problem-solving and logical thinking should come before putting so many efforts into learning new tools. When you get extremely good at problem-solving, your ability to learn and use tools needed to deliver will improve tremendously.
When I’m on a project, the majority of my time is spent on understanding and deconstructing the problem. After which I chart a logical path to achieving my aim. Upon achieving the two above and deciding on the analysis tools most suitable for the task at hand, I mostly just Google how to implement aspects of a tool I encounter problems. In Microsoft Excel, for instance, the only formulas I know at heart are Vlookup and If, the rest I use I just Google, copy and apply it to my work. I always know there is a formula for something I want to do or someone on the internet has encountered whatever problem I’m facing with a tool. But this confidence stems from previous understanding and deconstructing the problem into small bits. The same thing goes for tools such as Power Query and Python (pandas). In Pandas, for instance, many times, the libraries I had used on a project became known to me when I was charting the logical path to achieving my aim. Charting a logical path gives you insights into what tools and indeed aspects of those tools you need.
This is what many programmers do. They understand what they need to do and then just Google whenever they encounter issues. StackOverflow is heaven for programmers in this regard. In fact, many of them don’t encounter issues before going on StackOverflow to copy syntax. Why memorize syntax when you can get it easily?