Compute phi, bespoke percentiles, and evaluate interpolation error

[ddbuscombe]

let's compute and report all percentiles in millimeters, and make a quick notebook showing

• how to convert to phi from mm
• how to estimate interpolation error in mm
• derivation of other percentiles, e.g. d66 or whatever

[ddbuscombe]

I added a .gitignore to deal with the dataset (csv file), and added a way to download the data from google grive

I ran the code and got an error (related to inputs that are not explained), so...

• please break cells up, so each line of block (if loops) of code are in a separate cell
• add preamble / description of the purpose of the notebook, and how to use and modify
• show a complete example first, then show how a user may modify the example for their own purposes (make up 1 or 2 scenarios)
• explain your variables. No idea what "input number of given distributions in dataset' means - what would an average person with no prior background understand?
• add commentary throughout - what is this code doing?
• add details to README about this notebook - what does it do and why, and how to run it