According to udacity, these are essential skills every data scientist must have:
You are currently using the most popular and widespread programming language in the world. If you know how to use this package, this should qualify.
What is better than theoretical statistical skills? Practical applicable software modules. You know what you want out of your data, implementation details don't matter much.
TODO: pick ML libraries
TODO: pick Predictive libraries
- ordering data
- iterators for chunk modifications
- regular expressions
- date/time parsing
Developers mostly want JSON or CSV data, less technical people need excel files for tinkering and managers at most "read" rendered graphs, slides, or PDFs.
The more SE skills you have, the better a data scientist you will be. In smaller companies, there is a lot of groundwork to do, while in enterprises there are huge piles of messy data cluttered over many incompatible systems.
Deciding what makes sense - what to use and when - this is your actual job.