nmedina17 / osa Goto Github PK

start making new figs from updated plan in Results section.

stat plus graph

polish/adapt intro frmo previous Word verson.

while centers are needed for their own stats, for plots, add
variance to plot, maybe by plotting as stat_summary() instead of
direct calculated ceters from full tbl (i.e. $varData1).

map and mutate biomass data models and graphs. try to
select out timber sp onto graph.

maybe it will bring out
other variables as
predictors? but may explain
opposite relationship
between richness and
diversity. 

find out which
species cause the
diffs in
community/diversity

make histograms of size metrics
(DAP, height, biomass) to
supplement later diversity
metrics, which use instead species
frequencies to get entropy 

try summarizing
discrete distance
levels, since study
only analyzes 1
gradient at 1 study
site

general improvement/work

try making beta diversity among plots in addition or instead of ordination in fig2, suggested by Schloss
Riffomonas CC204.

choice of diversity metric from issue 6 still open, but for later. for now, re-model stats for non-normal
variables, and maybe integrate into big results table skeleton, if possible.

extracting pvals from nested summary tbls in col of main
stat tbl, then adding final col to eval it, per measured
var

basing choice on higher R2...or AIC?

try nlme to reduce "2" system
based on poly

organize current code into new file structure advised by Pat Schloss, remembering that the main notebook file is for discussion and presentation of most updated figures and numbers, and use separate files/scripts for each analysis--exploratory, or for each individual figure.

incorporate tidymodel::
skeleton into stat
pipeline?

wanting to try using 'tidyr::nest()' to run a basic linear
regression for all collected variables, but table structure
from 'rstatix::get_summary_stats()' is different (maybe already
a bit nested). solving this would really shorten the data
exploration phase, since the regression models would all
have the same independent variable--distance to forest.
potentially efficient!

with draft dispersal categories, run dispersal
analysis for fig3