If I had to give one piece of advice to early-career data scientists, it would be this:
✨ Practice asking questions until you understand the problem. ✨
The problem is usually not what you’re told. And this isn’t usually down to malice, but more that good managers are supposed to filter out noise and unnecessary complexity for you. In addition, they probably have some experience and ideas.
All this means that instead of the actual problem, people will helpfully offer what they think is the best approach.
Then your problem is simply implementing that approach.
This is actually the way to start out, and it can be a relief because it means you get to focus on the application of your newly acquired toolkit. But in the long run, it stops you from feeling fulfilled AND advancing in your career. Being able to independently design your own solutions and tackle harder problems requires you to understand how the nuances of a problem affect your approach.
Like everything else, it comes with practice.
Another point is this — unlike designing and building a bridge, data science is iterative.
Building a data science solution is like building a bridge when you’re working with some new cement that you haven’t fully figured out yet. You might become the one person in the room with the best understanding of the cement. That understanding needs to feed back into the design process.
The only way for that to happen effectively is if you have a good idea of what you’re trying to accomplish.
You’ll learn more, feel more satisfied, and have a bigger impact.