Skip to content

issue.md

Latest commit

 

History

History
56 lines (41 loc) · 1.92 KB

issue.md

File metadata and controls

56 lines (41 loc) · 1.92 KB
Raw

to be checked

data inputs:

long format with columns trait, abundance, species, (sites, coordinates in option) advantages: possibility to integrate intra-specific trait variation disadvantage: not the usual community matrix used in vegan or mobr

  • do we allow NA in abundance? or in traits?
  • no NA allowed in species, sites or coordinates when present
  • need easy function to transform data from vegan and mobr package
  • so far only one single trait, do we want to include multiple traits?

trasm new R-class

  • Species, Trait, Abundance : in that order
  • to be improved: add sites, coordinates and projection system
  • add conversion function from mobr and vegan

Functions

why scale_traits()

do we need it? does that change the estimates? SCALE_T? flat : for individuals, might be trickier ... should we take unique values of traits? unique combinaison of trait x species?

do we need VAR_T as indicator and ?

partition over m

covTA_viaTR = EMP - NULL_SAD ? pct_covTA_viaTR_signed = 100 * covTA_viaTR / denom_delta_signed, covTA_viaTR = EMP - NULL_SAD, # alternative path

make_steps()

  • not currently matching the number of steps because in log/sqrt scale it could be the same integer. Is it an issue? we should clarify it in the documentation
length(make_steps(m_use, n_points = 100, mode = step_scale))
  • naming: make_steps() is very similar to seq() in R, so maybe rename to seq_scale()

compute_curve_from_order()

  • renamed to get_curve_from_order
  • removed within_sd parameter
  • similar to vegan::rarefy() function
  • to be optimized?

summary_trasw

Should be weigthed indicators (weigthed mean and sd). same for beta_cov

how to compute rarefaction curves efficiently?

https://vegandevs.github.io/vegan/reference/rarefy.html https://vegandevs.github.io/vegan/reference/specaccum.html

specaccum test vs sa_all <- vegan::specaccum(wide) #, method="random"