Skip to content

Latest commit

 

History

History
68 lines (46 loc) · 3.17 KB

GRAPHVIZ.md

File metadata and controls

68 lines (46 loc) · 3.17 KB
Raw

General graph visualization

Graph decorators

The workhorse of the package is the decorator operator %@%. It accepts an igraph object (left side) and pipes it into a decorator function (right side). Decorator functions transform data into graph attributes and attach them to the correct slots in the graph object for use in the plot function. Several decorator functions in sequence can stack their outputs onto the graph. The pipeline ends with the magrittr pipe operator %>% to feed the decorated graph into the plot function:

g <- igraph::sample_gnm(n=10, m=15) %@%                      
     node_color(data=letters[1:10], palette='BuGn') %@%      
     node_size(data=1:10, min=5, max=8)                     

g %>% plot_netview()

Note that an alias for the graph decorator operator is %g% if you need to avoid the namespace clash with purrr::%@% which is an attribute assignment operator.

Decorator pipelines

Besides decorator functions, decorator operators can accept a data.frame or tibble containing the data to stack on the graph. This allows for passing the column name to the decorator function:

node_data <- tibble::tibble(x=letters[1:10], y=1:10)        # generate 10 x 2 node data tibble

g <- igraph::sample_gnm(n=10, m=15) %@%             # generate random graph with 10 nodes
     node_data %@%                                  # decorate graph with node data tibble
     node_color(data='x', palette='BuGn') %@%       # decorate nodes with colors paletted by x
     node_size(data='y', min=5, max=8)              # decorate nodes with values rescaled by y
     
g %>% plot_netview()                                # plot decorated graph from magrittr pipe

The pipeline can also be executed on a list of graphs:

graphs <- lapply(1:2, function(x) igraph::sample_gnm(n=10, m=15)) %@%
          node_data %@%                                                 
          node_color(data='x', palette='PuOr') %@%                      
          node_size(data='y', min=15, max=20)                          
     
graphs %>% plot_netview(legend='x')

Netview plots

The plot_netview function translates the decorated graphs into plots using igraph::plot.igraph. Note that all settings provided with decorator functions can be overwritten or refined by passing standard arguments for igraph::plot.igraph to netviewr::plot_netview. Graphs can be piped into the plot function using the magrittr operator %>% - this allows users to configure graphs by setting a basic configuration, and then change attributes on subsequent graph assignments.

base_graph <- igraph::sample_gnm(n=10, m=15) %@%
              node_data %@%                                      
              node_size(data='y', min=5, max=8) 

base_graph %@% node_color(data='x', palette='BuGn') %>% plot_netview()
base_graph %@% node_color(data='x', palette='PuBu') %>% plot_netview()

Legends and titles can be set using the legend and text arguments. Legend parameter is character that selects a column from attached node data when %@% is used (see above)

graphs %>% plot_netview(legend='x', text="Test", text_size=1.5)