Ability #1 == to translate "human questions" into bite-size hypotheses. By "human questions," I mean conceptual and high-level questions that start data science projects (e.g., Is our company diverse and inclusive? What additional services would help our current customers?) A data scientist should be able to break down a human question and convert it into hypotheses that are testable. This task is not possible without a great understanding of the "human question" (where it is coming from and what impact the answers would have, etc) as well as the data (what is measured, how it is collected, etc).
Ability #2 == to fine the right tool for analysis and to code.
Ability #3 == to translate back the results of the analyses into human language. In other words, storytelling.
When I just got interested in data science, all I thought about was coding. It's about BIG data and you need to know how to deal with them. Right? Now I know that data science is about problem solving--coding just provides a powerful tool.
It is relatively easy and straightforward to acquire Ability #2. There are many courses and resources out there. However, no one could "teach" you Ability #1 and #3. Honing them require significant time, experiences, brain power, and curiosity. And they separate great data scientists from good data scientists.
