Node Types & Styling

DAGMakie provides semantic node types for causal diagrams, with automatic styling.

Node Types

The NodeType enum defines common roles in causal diagrams:

@enum NodeType begin
    Observed      # Standard observed variable
    Latent        # Unobserved/latent variable
    Treatment     # Treatment/exposure (thick stroke)
    Outcome       # Outcome (emphasised outline)
    Instrument    # Instrumental variable
    Confounder    # Confounding variable
    Mediator      # Mediating variable
    Collider      # Collider variable
    EffectMeasure # IDAG causal-effect node
    SwigFixed     # Fixed half of a SWIG split
end

For interaction / DiD conventions (IDAGs, modifier edges, SWIGs — defined there), see Visual Grammar.

Using DAGSpec

The DAGSpec type combines a graph with node specifications:

using Graphs, DAGMakie, CairoMakie

g = SimpleDiGraph(3)
add_edge!(g, 1, 2)  # Z → X
add_edge!(g, 1, 3)  # Z → Y
add_edge!(g, 2, 3)  # X → Y

spec = DAGSpec(g,
    node_labels = ["Z", "X", "Y"],
    node_types = [Confounder, Treatment, Outcome]
)

# Auto triangle layout: confounder Z at the apex
fig, ax, p = dagplot(spec)
fig

Node Type Styling

Each node type has default styling (in-node white labels on dark fills, except Latent / SwigFixed):

Type	Default Colour	Marker / stroke	Use Case
`Observed`	Steel blue	circle	Standard observed variables
`Latent`	Gray (hollow)	circle, stroke 2	Unobserved variables
`Treatment`	Steel blue	circle, stroke 2.5	Exposure / intervention
`Outcome`	Steel blue	circle, stroke 2.0 darkgray	Response variables
`Instrument`	Goldenrod	diamond	Instrumental variables
`Confounder`	Goldenrod	circle	Confounding variables
`Mediator`	Seagreen	circle	Mediating variables
`Collider`	Medium purple	circle	Collider variables
`EffectMeasure`	Seagreen	rect	IDAG effect-measure node
`SwigFixed`	White	rect, stroke 2	SWIG fixed half $a$

Custom Styling Functions

Override default styling:

# Get default colour for a type
color = default_node_color(Treatment)  # :steelblue

# Get default marker
marker = default_node_marker(EffectMeasure)  # :rect

# Get default stroke width
stroke = default_node_strokewidth(Treatment)  # 2.5

Apply Node Type Styling

Apply styling to existing graphs:

g = SimpleDiGraph(3)
add_edge!(g, 1, 2)
add_edge!(g, 2, 3)

types = [Confounder, Treatment, Outcome]
colors, markers, strokewidths = apply_node_type_styling(types)

fig, ax, p = dagplot(g,
    nlabels = ["Z", "X", "Y"],
    node_color = colors,
    node_marker = markers,
    node_strokewidth = strokewidths,
)
fig

Pre-typed Graph Constructors

Create graphs with pre-assigned node types:

spec = typed_confounding_graph()
fig, ax, p = dagplot(spec)
fig

fig = Figure(size = (900, 240))
ax1 = Axis(fig[1, 1], title = "typed_mediation")
ax2 = Axis(fig[1, 2], title = "typed_instrumental")
ax3 = Axis(fig[1, 3], title = "typed_collider")
dagplot!(ax1, typed_mediation_graph())
dagplot!(ax2, typed_instrumental_graph())
dagplot!(ax3, typed_collider_graph())
fig

Style Presets

DAGMakie provides style presets for different contexts:

g, labels = confounding_graph(["Z", "X", "Y"])
fig = Figure(size = (900, 240))
ax1 = Axis(fig[1, 1], title = "default_style")
ax2 = Axis(fig[1, 2], title = "minimal_style")
ax3 = Axis(fig[1, 3], title = "bold_style")
ax4 = Axis(fig[1, 4], title = "presentation_style")
dagplot!(ax1, g; nlabels = labels, style = default_style())
dagplot!(ax2, g; nlabels = labels, style = minimal_style())
dagplot!(ax3, g; nlabels = labels, style = bold_style())
dagplot!(ax4, g; nlabels = labels, style = presentation_style())
fig