search

tesselle / tabula

:bar_chart: Analysis and Visualization of Archaeological Count Data
https://packages.tesselle.org/tabula/
GNU General Public License v3.0
35 stars 5 forks source link

JOSS #15

Closed soodoku closed 4 years ago

soodoku commented 4 years ago

Hey @nfrerebeau: I have created a few issues. when you get a chance take a look. The key thing that is missing in the paper =

"State of the field: Do the authors describe how this software compares to other commonly-used packages?" Specifically, can you speak to all the other packages tailored toward ecologists etc. like https://github.com/vegandevs/vegan (I don't know as much about the field)

For the package, the key thing that is concerning = re-implementation of classic but simple algorithms, like shannon diversity, etc. it may be useful to import robust implementations? I don't know if they exist so suspect that they do.

The key thing missing from the package = whenever stat. tests are done, there is insufficient detail in the man pages about the assumptions and what precisely happened. There are some cites. But useful to write in a more expanded fashion what was done and return an object that has, for instance, all the estimates from each of the bootstrapped sample.

Please disagree with whatever doesn't make sense. I don't want you to do anything that is not adding value. My aim is just to help make the package better.

nfrerebeau commented 4 years ago

For the package, the key thing that is concerning = re-implementation of classic but simple algorithms, like shannon diversity, etc. it may be useful to import robust implementations? I don't know if they exist so suspect that they do.

I have added three new methods (index_heterogeneity, index_evenness and index_richness - diversity, evenness and richness are now deprecated) that return the diversity index estimates as well as the jackknifed and bootstrapped statistics. These estimates can be compared to simulated assemblages and plotted for graphical inspection.

soodoku commented 4 years ago

cool.

tbh, my point didn't require as much work. i was just saying that some other people may have already produced robust implementations. by robust, i narrowly meant things that catch the run of the mill errors and prevent things from exploding etc. --- ala log 0, etc.

but this looks good. i still need to check the functionality but it looks good to me.

nfrerebeau commented 4 years ago

"State of the field: Do the authors describe how this software compares to other commonly-used packages?" Specifically, can you speak to all the other packages tailored toward ecologists etc. like https://github.com/vegandevs/vegan (I don't know as much about the field)

I have added a paragraph to clarify this point.

soodoku commented 4 years ago

super. feel free to close it out.