I am currently working on creating figures for my manuscript - Due to the design of the study I chose to use GLMMs. However, I am now struggling to find some codes on how to illustrate my results. I am measuring the effects of forest degradation (continues variables) on community structure with sites as random effect.
I have used ggplot2 on my first draft, but I was advised from a reviewer to show some partial plots and also to plot the model together rather than use only singular linear regressions .
Any help would be greatly appreciated.
"Nonlinear Regression and Generalized Linear Models" https://biol607.github.io/lab/12_gzlm.html
Here's one example of what you might do. Best wishes, David Booth
To add to the excellent suggestions, two things am going to add:
1. On codes to illustrate the GLMM results (multiple regression models, including partial or split graphs. The package "mvabund" is one of the many complements to R graphics, lattice and ggplot2 that produces plots for the visualisation of multivariate abundance data and their relationships to environmental variables. Contour, heatmap and joint boxplots are good, as well as chlorophet maps for the forest areas/sies. See https://rpubs.com/dwarton/68823 for motivation and related samples. There are many tutorials, Google it pls, or mail sample data and regression to inbox for specific coding.
2. On combining the partial and full graphics, it's a matter of what format or graphics are really needed (best capture the analysis information) and how you are able to use the documentation/tutorials. R base graphics placement control function par(mfrow=c(r,c)) can be used to specify either the number of rows or columns the graphics will occupy or just sequentially plot the models (mod1, mod2, etc). Perhaps, the easiest approach to assemble multiple plots include using the facet_wrap() and facet_grid() function of the ggplot2 package (when you are creating multiple plots and they share axes). i.e., write your ggplot2 code as if you were putting all of the data onto one plot, and then you use one of the faceting functions to specify how to slice up the graph;
Another is the grid.arrange() function from the gridExtra package, fir example 👉grid.arrange(p1, p2, nrow = 1).
Finally, you can use also use the cowplot package function, plot_grid(), to combine the two plots of your interest. And, plot multiple models using library(sjPlot)'s function plot_models(fit1, fit2, fit3, grid = TRUE)
EXTRAS:
* par(mar=c(2,10,2,2)) # adjusts the graphic window margins.
boxplot(Degradation$sites, horizontal = TRUE, las=2, main="Effects of Forest Degradation on Community Structure, Given Sites")
**And, consider "ggplot2 - Easy Way to Mix Multiple Graphs on The Same Page" at http://www.sthda.com/english/articles/24-ggpubr-publication-ready-plots/81-ggplot2-easy-way-to-mix-multiple-graphs-on-the-same-page/
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators