The following repository contains data and code used for analyses in:
Tran, N.-H., Atmaca, S., Waring, T. & Beheim, B. A. (2021). Entropy trade-offs in artistic design: A case study of Tamil kolam.
These instructions will get you a copy of the project up and running on your local machine.
In order to be able to run the code, you will need to have an up-to-date version of R installed on your computer and a few CRAN packages (see below). You will need Stan in order to fit the statistical models. If your operating system is macOS Catalina or higher, you might encounter C++ compiler problems with Stan. To fix these C++ compiler problems, please check the following links: Link1 and Link2. To be able to plot the kolam drawings, you will further need to install the kolam R package.
# Install devtools package if necessary
if(!"devtools" %in% rownames(installed.packages())) install.packages("devtools")
library(devtools)
Sys.setenv("R_REMOTES_NO_ERRORS_FROM_WARNINGS"=TRUE)
devtools::install_github("rmcelreath/rethinking")
# Install required packages
install.packages(c("rstan", "ggpubr", "dplyr", "msm", "ggplot2", "ggdag", "parallel", "tictoc", "patchwork", "ggridges", "ks", "concaveman", "igraph", "tidyverse", "Rsolnp", "gtools"))
# Install the stable development version of the kolam package
devtools::install_github("nhtran93/kolam")
# Install the stable development of kernelgon to create Figure 4 in the manuscript
Sys.setenv("R_REMOTES_NO_ERRORS_FROM_WARNINGS"=TRUE)
devtools::install_github("babeheim/kernelgon")
The data to run the analyses can be found all in data/kolam_data_share.RData or in form of a .csv file. To get the full sample object for each model, you will have to fit the statistical models.
data/: Contains the data.
output/: Figures and posterior samples will be stored in this directory.
helpers/: Contains all the helper functions for the analyses.
stan/: Contains the Stan code for the random intercept model.
samples/: This directory will store the MCMC samples.
Download all files from https://github.com/nhtran93/kolam_signaling
If you use R, please set the working directory to the appropriate directory where you have saved these files. If you use RStudio, you can just open the respective RStudio project and the directory will be automatically set.
Some example kolam transcription files are in the repository in form of .yaml files. Code for all the figures in the manuscript can be found here.
Before fitting the statistical models, you will first need to check your directory:
# Make sure to check your working directory first. You should be in the root directory of this project.
getwd()
# Make sure to create a directory where the fitted MCMC samples will be saved.
dir.create("samples")
# Example of how to fit a statistical model (change the if statement to TRUE):
density_model <- FALSE
gini_model <- TRUE # This model will be fitted
unique_model <- FALSE
canvas_size_model <- FALSE
entropy_model <- FALSE
IMPORTANT NOTE: Make sure you have enough disk space and working memory left on your computer. The statistical models take a while (~2h each) to run and the samples are each up to 2GB big. It is recommended to run the statistical models on a cluster computer if possible.
You can run fit the statistical models one by one using the R script Fit_Models_intercept.R to run the random incetercept models. The code chunks to fit the statistical models are each wrapped in an if statement. The default of the if statement is FALSE. If you want to fit the statistical model, you will have to manually change the if statement into TRUE on the top of the R script. All your fitted MCMC samples will be saved in the directory samples. Afterwards, you can describe the models and run diagnostic plots (Figures shown in the main paper and in the SI). The plots will be automatically saved in the directory output. Please note that the R script Describe_model_samples.R will not be able to produce the plots without previously fitting all the statistical models.
You can further find the code to run the simulations to solve for the entropy envelope in terms of boundary conditions and reproduce the SI Figure 1 in entropy_evenness_richness_simulation.R.
This project is licensed under the CC-BY-4.0 License - see the LICENSE.md file for details.