If I had to design a data science curriculum, I would try to cover essentials and open doors to further study without overloading the student.

The field is so huge and as-yet-undefined, that skill sets and experience can be hard to fit into a nice neat definition.

That said, I’d aim for something like this:

• linear algebra 1st/2nd year uni
• statistics 1st/2nd year uni
• probability theory 1st/2nd year uni
• programming i.e. writing good code and a bit of application design, but not full blown software engineering
• relational/non relational DB design, modelling and SQL
• a language, dependent on students’ aims, field of interest, previous experience. Obvious and common candidates are python, R, matlab and SAS.
  • Yes SAS is big. Especially in medicine, government and pharma. It’s not going away.
• lin/log regress, neural nets, random forests
  • point the way for them to study ML deeper
• data visualisation, based on the work of Tufte, Wickham, Munzner

And that’s probably it. My master’s degree had this and more, and this was on top of a science degree, plus my own study of mathematics and computer science.

Traditional maths education usually requires some calculus before linear algebra and probability - I’m still of two minds as to whether this is still really necessary.

The above short list should be enough for a diploma level qualification and should help immensely in any role or organisation.