Skip to contents

Scatter plot of eigenvector centrality vs dMGI

Source: R/gravity_plots.R
plot_gravity_scatter.Rd

Plots eigenvector centrality on the x-axis against net gravity (dMGI) on the y-axis for all modality nodes. Points are coloured by role (attractor/satellite/neutral) and contradiction cases - nodes where the two metrics disagree most strongly - are automatically labelled. The Spearman correlation is annotated in the plot margin.

Usage

plot_gravity_scatter(
  x,
  catmodgraph,
  label_threshold = 0.25,
  attractor_col = "#1D9E75",
  satellite_col = "#D85A30",
  neutral_col = "grey60",
  point_size = 1.8
)

Arguments

x

A data frame returned by modality_gravity.

catmodgraph

A catmodgraph object matching x. Used to compute eigenvector centrality if not already present as a column in x.

label_threshold

Numeric in \[0, 1\]. Modalities are labelled if their eigenvector centrality exceeds this value and they are satellites, OR if their delta_mgi exceeds the 75th percentile and their eigenvector is below the median. Default 0.25.

attractor_col

Colour for attractor nodes. Default "#1D9E75".

satellite_col

Colour for satellite nodes. Default "#D85A30".

neutral_col

Colour for neutral nodes. Default "grey60".

point_size

Numeric. Base point size. Default 1.8.

Value

Invisibly returns a data frame with columns node, eigenvec, delta_mgi, role, and is_contradiction.

Details

This plot is the primary diagnostic for demonstrating that MGI captures structural information not contained in standard centrality indices.

See also

modality_gravity, plot_gravity

Examples

data(survey_health)
mg   <- build_modality_graph(survey_health)
mg   <- prune_modality_edges(mg, min_weight = 0.10, max_p = 0.05)
grav <- modality_gravity(mg)
plot_gravity_scatter(grav, mg)