forestgeo / forum Goto Github PK

From @maurolepore on July 18, 2017 20:45

Code for cdarbon analyses may be packaged, either inside forestr or, if it makes sense, as a stand alone module.

"The only way you can intellectually manage a large program is to break it into pieces so that you only have to think about one part at a time. Modularization is the most powerful tool at your disposal for breaking a program into pieces."
--"Code Complete (Developer Best Practices)" by Steve McConnell http://a.co/btQnbMX

This needs Helene Muller-Landau [email protected] and Sturat to agree.

Copied from original issue: forestgeo/fgeo.abundance#25

@mcgregorian1 and @teixeirak,

Cara (@crscalpone) is working for Sturat and some of her code may be a good fit for your plant-book.
Feel free to dicsuss here and see if you want to invite her to your private repo to continue the discussion there.

I'm looking for volunteers to test the installation process of the fgeo package. Your feedback will make fgeo easier to use and benefit the ForestGEO commumity.

• Could you install fgeo? Which installation option (1 or 2) did you use?
• What problems did you experience?
• What was unclear/confusing?

Please let me know any issue, comment, or idea for improvement. Write below or at [email protected].

cc': @rutujact, @crscalpone, @kccushman, @ervanSTRI, @jess-shue, @DanielZuleta, @ngokangmin, @gonzalezeb, @dkenfack, @ekraichak, @forestgeoguest, @mopicon, @ValentineHerr

A workshop at the 2017 meeting of the British Ecological Society is calling for challenges.

Suggestions for challenges should ideally revolve around the following topics:
· interfacing ecological databases with R;
· combining different databases / data sources;
· advancing statistical methods;
· creating tutorials for ecological analyses.

https://docs.google.com/document/d/1JiLJx_kM48JXqjiUiZh4cJEm5UNCfMnNijly40roS5w/mobilebasic?usp=gmail

This might be a good idea:

"[It would be useful to] Create an R package with a collection of well-documented ecological datasets ... similar to agridat"

---Florian Hartig

For inspiration, see these public data-repositories by @teixeirak:

• https://github.com/forestgeo/Climate
• https://github.com/forestgeo/Site-Data

Proposed tutorials and people interested are listed here: here.

• Create forum,

• Provide instructions to participate.

Does anyone have code to calculate the abundance of all species withing a radius?

(On behalf of Basset, Yves [email protected])

In the CTFS R Package (http://ctfs.si.edu/Public/CTFSRPackage/) the function
NeighborDensities() is close but not exactly what Ives wants.

Suzanne Lao run NeighborDensities() for the species quaras:

one.sp = subset(bci.full7,  sp == "quaras")
neighbor.counts <- NeighborDensities(bci.full7, one.sp, type='count')

The output gives the number of conspecific trees and the number of heterospecific trees within a radius (20 m by default) of each tree of the quaras species. But Yves wants to calculate the abundance of all species. Does anyone have code to calculate that?

Please contact Basset, Yves [email protected] and copy to [email protected].

From @maurolepore on July 18, 2017 18:29

Matteo Detto suggested at the workhop in Puerto Rico (2017) that we could contribute some functions. Follow this up.

The code needs to be translated from MATLAB to R, but Matteo suggested the functions are small.

Copied from original issue: forestgeo/fgeo.abundance#22

From @maurolepore on July 18, 2017 18:27

?wavelet.allsp (and friends)

Copied from original issue: forestgeo/fgeo.abundance#21

The people listed below still need to accept our invitation.

Please check your email to accept the invitation to the ForestGEO's organization on GitHub. If you prefer, come and see me; I'll help you to set this up and to manage GitHub's notifications.

@Caicai112
@dkenfack
@EMGora
@kkkliuheming
@MariaNatalia
@pavel-fibich
@renatovalenciar
@sjosephwright
@tzeleongyao

It may be worth to think about error propagation using census data. There are some packages that compute error propagation using info stored in attributes.

https://github.com/Enchufa2/errors

From @maurolepore on March 9, 2018 18:54

The Smithsonian Institution has recently partnered with The Carpentries. The Carpentries -- Software Carpentry and Data Carpentry -- help researchers worldwide to become more productive by providing free training on data and software management.

Two Data Carpentry workshops have been delived to date (March 2018), one at NMNH and one at SERC. At SERC, for example, the workshop covered this content.

Two more workshops are expected during 2018. If you would like to participate, keep alert because once registration opens the available slots (~30-40) tend to fill extremely fast.

Copied from original issue: forestgeo/learn#63

#Hi @ervanSTRI and all users of data.table,

If you write functions with the package data.table, you may need to learn a bit about non-standard evaluation; or you may get unexpected or silent errors. To learn generally about non-standard evaluation, I recommend this: http://adv-r.had.co.nz/Computing-on-the-language.html

data.table, at least the by argument, seems to use non-standard evaluation. Non-standard evaluation is great when you work interactively; but it can bite you if you use it inside the functions you write. In the base case scenario, the problem will be that you get an unexpected error. In the worse case scenario, the error will be silent: you will get output but it is wrong.

I am not yet familiar with data.table, and the best I can do right now is to point you to examples of
non-standard evaluation in the package dplyr.

dplyr also uses non-standard evaluation; and provides an approach to writing functions with dplyr that is called tidy eval: http://dplyr.tidyverse.org/articles/programming.html.

From @maurolepore on July 18, 2017 20:5

dan johnson ([email protected]) has code in a package that was build from the CTFS R Package. Benchmarking. It would be good to see how that can be integrated.

• that script has a summay for any plot.

Copied from original issue: forestgeo/fgeo.abundance#24

Check for option to set frequency of eamils

The current biomass calculations use older tropical allometries. I am sure others have updated how they are calculating biomass on our plots. I think there should be a focused discussion on how to best keep the biomass function flexible (geographic variation in allometry) and current.

To all those who are following the development of fgeo,

I would like to end the year thanking you for your support and announcing the first candidate for a widely-announced release of fgeo. Before the release, I would love you to give me some feedback, and report issues.

Happy new year!
Mauro

Dear Stuart & Mauro,

I am currently wrapping up my code and I am facing difficulties to ensure that it is generic enough to be applied across different sites. Sure it's hard to avoid peculiarities among sites, but it might be worth thinking about possible ways to homogenize functions (i.e. having consistent site name, mnemonic, variables names) and somehow propose some "coding" guideline to potential GIThub users.
My 2 cents.
Ervan