Tom Slater's peer review: model inputs

[amyheather]

Peer review from @tbslater

Model Inputs

Input Modelling

• I would put the functions on a new line following the pipe operator for clarity (like you did for the inter-arrival times).

• I would include a short description of the code. Even something as simple as the function below is used for producing the time series plot, given the data and columns to plot.

• Some example plots would be useful here.

• It might be worth stating the null and alternative hypotheses for the KS test, as a larger p-value being ‘better’ in some sense is rather unintuitive (I’m thinking about other cases where rejecting a null hypothesis, and thus a small p-value, is desirable).

• I think this subsection would benefit from a ‘cheat sheet’ of input modelling distributions with some plots of their general shape. I think it may be unclear for people unfamiliar with statistical distributions why you have chosen certain ones.

Input Data Management

• Typo...

The checklists are great!

Parameters from Script

I really liked this subsection. I particularly like how you build up from ‘what not to do’ to the preferred approach – I think that helps convey the reasons for grouping and passing. Everything was super clear and well-presented.

Parameters from File

Again – very clear.

• Where should the parameter file be stored if not structured as a package? Does it matter? Or is part of the overall aim to encourage people to structure as a package?

Parameter Validation

No comments here 😊