search

andrew-edwards / sizeSpectra

R package for fitting size spectra to ecological data (including binned data)
10 stars 6 forks source link

Reorder README to better explain how users can fit size spectra to data, rather than reproducibility of methods #9

Open jpwrobinson opened 1 year ago

jpwrobinson commented 1 year ago

Currently the first impression from README is this is a supplementary repo for the MEE/MEPS papers, rather than a package that people can use for their own analyses.

Simple fix would be to bring user information above the details, e.g.

Short intro - “SizeSpectra package provides code for estimating size spectrum exponent of body size data, such as collected on fish assemblages for fisheries monitoring purposes. The code developed by AE and explained in MEE + MEPS papers.”

1) Installation 2) New size spectrum function - how to fit MLE to my data 3) VIgnettes 4) Citation + Issues 5) Background on papers 6) Updates

andrew-edwards commented 1 year ago

Thanks - great to chat about this the other day. Yes, I'll try and take a stab at it soon. People have used the package a lot, but, yes, I can make a simpler function (I think for MLEbin it's easier, I just never went back to point out how to do a simple MLE fit).

andrew-edwards commented 1 year ago

@jpwrobinson - I haven't forgotten about this, just had no time yet.

I'm now thinking of starting a new user-friendly package from scratch, keeping this one to be the reproducing of the results from both papers. The current package was adapted from code that had built up over many years, and has some inconsistent variable names etc. in there, and clunky functionality. I'm thinking it will be easier to start again (obviously using some existing code as a template), to ensure consistency and allow proper automatic unit testing etc. which is very useful for improving a package that people are using. I'd probably call it sizeSpectra2 I expect!

Will realistically have to be in 2024 some time, but I've written a few package lately so can envision how to do it.... Thanks for the motivation.

andrew-edwards commented 2 months ago

@jwprobinson - I have started setting up a new streamlined package! Going with calling it sizeSpectra2, though wondering if sizeSpectrum is better (since it will be focussed on MLE methods, producing just a single size spectrum for a given data set, not 8 like in our simulation papers). Any thoughts?