distributing data to students

[orbeckst]

What is the best way to provide the planet data to students?

• just as a table and they have to come up with data structures
• CSV file (and tell them about CSV readers)
• JSON (import json) but need to make sure that floating point numbers are stored accurately
• hdf5 and use pandas...
• python file where we store everything as a dict and they import it?

Other ideas? What will work best – @ianmkenney ?

(@dotsdl – also interested in your opinion: internally we used data_manipulation.ipynb to prepare data and in my draft of the solution p2_solution.ipynb I've been liberally using pandas but the students do not know pandas, only numpy)

[ianmkenney]

I see nothing wrong with just providing the dict. However, we could make a point by providing a YAML file with instructions on how to get something like pyyaml via pip. From there we can even provide the parsing code.

[orbeckst]

Done – distributed dicts in parameters.py.