Skip to content

Latest commit

 

History

History
89 lines (57 loc) · 4.04 KB

DECFUNC.md

File metadata and controls

89 lines (57 loc) · 4.04 KB
Raw

Decorator functions

Decorator functions modify the graph object by translating data into graph attributes and attaching it to the appropriate slots in the graph object. These are eventually recognized by the plot_netview function so that graphs can be constructed and decorated independently of the igraph::plot.igraph arguments, which can be a little arcane in their original implementation.

Node decorators

node_color(g, data=NULL, condition=NULL, palette='BuGn', color='gray', opacity=1, n_color=NULL)

Map string or factorial data to node colors from a palette or set a uniform color if no data is provided.

node_size(g, data=NULL, size=4,  min=5, max=10)

Map numeric data to node sizes, rescale the data to a range of min and max sizes, or set a uniform size if no data is provided.

node_label(g, data=NULL, label=NA, size=0.8, color='black', family='serif', font=1, dist=0)

Map string data as labels or set a uniform label if no data is provided. Labels can be somewhat refined with arguments in this decorator, but more sophisticated options are available in igraph::plot.igraph.

node_shape(g, data=NULL, shape=c('circle', 'square', ...))

Map string or factorial data to a node shape, recycles values if more data levels available than shapes.

Special decorators

node_pie(g, data=NULL, palette='BuGn', n_color=NULL, border_color='black', match_func=dplyr::starts_with)

Map proportional data values over m categories (must sum to 1) to nodes that represent pie graphs, where each slice is colored according to palette. Data can be a tibble, data.frame or string column name, in which case the data argument uses the match_func to select the columns from the data the graph was decorated with. For instance, if all columns that hold the pie graph values start with pie, such as pie_1, pie_2 and pie_3 the data argument would be data = 'pie'.

community(g, method='walktrap', polygon=NULL, palette='PuBuGn', opacity=1, border=NA, n_color=NULL, ...)

Detect communities in the graph and draw polygons colored by palette around the clusters in the network topology. Multiple community detection methods can be specified. If polygon is a boolean and TRUE the first method is used to draw polygons, otherwise the polygon accepts a string specifying which community detections results should be drawn, e.g. polygon = 'walktrap'. Any other named arguments that specify the behaviour of the community detection algorithm can be passed via the ellipsis argument (...). Community detection can be run before assigning other node attributes. The name of the community detection algorithm then maps community membership to node attributes, for example:

g <- igraph::sample_gnm(n=10, m=15) %@% 
     community(method='walktrap') %@%
     node_color(data='walktrap', palette='BuGn')

Custom decorators

Decorators are functions that require a graph object (g) and return an anonymous function (func) that first modifies and then returns the graph object. A simple example would be a decorator that assigns node colors via a colorize function. Without error checks and other essential behaviour and for demonstration purpose only, this could then be written as:

node_color <- function(g, data=NULL, palette='BuGn', ...){
  func <- (g, ...) {
    V(g)$color <- colorize(data, palette)
    return(g)  
  }
  return(func)
}

Geonet

Geographical decorators and projections under development