wpetry / structuralcoexistence Goto Github PK

If either the "Show table" or the "Show network" checkboxes are un-selected by the user, both the table and the network disappear:

Image 1: both shown

Image 2: "Show table" unselected

Image 3: "Show network" unselected

(both unselected works as expected)

Is there a way to add in a real data set, e.g. the Sedgwick species used in the original paper? The idea would be to support inquiry-based activities. Users would select species from 3 (or 4) grouped dropdown menus, and their corresponding r's and alphas would fill in the parameter boxes.

graphical output should be rotatable by the user

maybe one of these?
https://trestletech.com/2013/10/interactive-3d-in-shiny-shinyrgl/
https://stackoverflow.com/questions/31327076/spinning-3d-scatterplots-inside-a-shiny-app

From @Granjel:

I have gone through the code of your StructuralCoexistence shiny app (congrats, by the way, it's really useful) and, by doing so, I've noticed that you transform the structural niche differences assuming a natural logarithm and not a decimal one, after calculation with the function Omega() from Saavedra et al. 2017.
I had this doubt about a year ago and I asked Serguei directly. In the end, we agreed that the proper transformation should be 10^x and not exp(x). So, to really see what is happening, I've made a little plot (attached) showing the differences between the structural niche differences calculated with the compute_overlap() function and both transformations of the Omega() function: the decimal transformation and the natural transformation. The blue dashed lines in the graph show the 1:1 line while the grey solid line is fitted by a simple linear model for the results of 1000 random 3 by 3 matrices. It seems that, by using the natural transformation, you might be overestimating the structural niche differences.

The visuals should be spruced up/made more consistent between figures within and among the 3-/4-species cases.

• consistency of section headers/labels on the left side menu
• static vs. interactive cone plots
• annotate structural niche/fitness differences on ternary plots
• coloring species nodes
  • in interaction network
  • labels in ternary/simplex plot
  • blend colors to show feasibility domains for species pairs in ternary plot?
  • add single species color to background of ternary plot?

Ideally it would be nice to have them independently turned on and off to enable clarity in teaching (i.e., building up to quantitative measures after visual intuition is developed).

allow user to select species, retrieving data from Figshare repo

Add present parallel scenarios where possible. Rock-paper-scissors becomes rock-paper-scissors-[lizard or Spock].