In jameelalsalam/nestedcats: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

To represent the potentially complicated nesting of categories, we will use a graph. In particular, the graph must be a tree or directed acyclic graph (DAG) such that the edges represent subset relationships.

The machinery for working with trees and graphs can be a bit unfamiliar (especially to me) so this article explores the different ways that the graph data structure can be constructed and represented inside of nested categories and/or partially ordered factor objects.

suppressPackageStartupMessages(library(tidyverse))
suppressPackageStartupMessages(library(igraph))
library(nestedcats)

ghg_nodes_df <- read_csv(system.file("extdata", "ghg_cats_nodes.csv", package = "nestedcats"))
ghg_edges_df <- read_csv(system.file("extdata", "ghg_cats_edges.csv", package = "nestedcats"))

ghg_edges_df

The correct format for the igraph::graph_from_data_frame constructor is a two-column data frame, with symbolic node names in each, and then a second data frame with node attributes. That is, the edge_df can have integer or character values.

ghg_graph <- graph_from_data_frame(ghg_edges_df, vertices = ghg_nodes_df)
ghg_graph

The key features are that:

• the secondary vertices data frame have the same list of node names as are used in the first two columns of the edge df.
• the order of the secondary vertices df becomes the underlying node numbers in the graph.

jameelalsalam/nestedcats documentation built on June 2, 2020, 8:16 p.m.