Graphs

NetworkX graph construction, analysis and drawing.

10 nodes. Right-click any node in the editor to read this documentation in the app.

Graph

🎯 Graph Centrality

id graph_centrality · Graph · Python export: yes

Score every node by degree, betweenness, closeness, eigenvector centrality or PageRank, highest first. Wire the result into a Graph Plot to size or colour nodes by it.

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).

Outputs

Port	Type	Description
centrality	DataFrame	Columns: node, <measure>; sorted highest first.

Fields

Field	Type	Default	Choices
Measure	select	degree	degree, betweenness, closeness, eigenvector, pagerank

🔲 Graph from Adjacency Matrix

id graph_from_adjacency · Graph · Python export: yes

Build a NetworkX graph from a square adjacency-matrix DataFrame (rows and columns are the nodes, cell values the edge weights). Weights at or below the threshold are dropped.

Inputs

Port	Type	Description
dataframe	DataFrame	Square adjacency matrix.

Outputs

Port	Type	Description
graph	networkx.Graph	The graph (DiGraph if Directed).

Fields

Field	Type	Default	Choices
Directed	checkbox	false
Edge threshold (|value| >)	float	0.0

🕸️ Graph from Edge List

id graph_from_edgelist · Graph · Python export: yes

Build a NetworkX graph from a DataFrame with one row per edge: name the source and target columns, optionally a weight column, and tick Directed for a directed graph.

Inputs

Port	Type	Description
dataframe	DataFrame	Edge list: one row per edge.

Outputs

Port	Type	Description
graph	networkx.Graph	The graph (DiGraph if Directed).

Fields

Field	Type	Default	Choices
Source column	text	source
Target column	text	target
Weight column (optional)	text
Directed	checkbox	false

🕸️ Graph Plot (Interactive)

id graph_plot_interactive · Graph · Python export: yes

Draggable, zoomable Cytoscape.js view of a graph with several layouts. Colour and label nodes and edges by any attribute (handy for knowledge graphs) and size nodes by a centrality column. Right- click the preview for full screen. In exported Python this node is skipped (the graph passes through) — use Graph Plot (Static) to draw the network in a script.

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).
centrality	DataFrame	Optional: output of Graph Centrality, to size / colour nodes by.

Outputs

Port	Type	Description
graph	networkx.Graph	The input graph, unchanged.

Fields

Field	Type	Default	Choices
Layout	select	cose	cose, breadthfirst, circle, concentric, grid, random
Node label attribute	text
Node group/type attribute (colour)	text
Edge label attribute	text
Size by (centrality column)	text
Title	text	Graph

🔗 Graph Plot (Static)

id graph_plot_static · Graph · Python export: yes

Plotly network plot of a graph using a networkx layout (spring, Kamada-Kawai, circular, shell, random). Size or colour nodes by a centrality column. Less interactive than the Cytoscape view, but it exports to Python.

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).
centrality	DataFrame	Optional: output of Graph Centrality, to size / colour nodes by.

Outputs

Port	Type	Description
graph	networkx.Graph	The input graph, unchanged.

Fields

Field	Type	Default	Choices
Layout	select	spring	spring, kamada_kawai, circular, shell, random
Size by (centrality column)	text
Colour by (centrality column)	text
Title	text	Graph

📊 Graph Summary

id graph_summary · Graph · Python export: yes

One-row table of headline statistics: node and edge counts, density, connectivity, number of components, average clustering and diameter (skipped for disconnected or 500+ node graphs).

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).

Outputs

Port	Type	Description
summary	DataFrame	A single row of graph statistics.

🔳 Graph to Adjacency Matrix

id graph_to_adjacency · Graph · Python export: yes

Turn a graph into an adjacency-matrix DataFrame (edge weights, 0 where there is no edge).

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).

Outputs

Port	Type	Description
matrix	DataFrame	Square matrix, rows and columns are the nodes.

📋 Graph to Edge List

id graph_to_edgelist · Graph · Python export: yes

Turn a graph back into an edge-list DataFrame (source, target and any edge attributes).

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).

Outputs

Port	Type	Description
edges	DataFrame	One row per edge.

🏷️ Set Edge Attributes

id graph_set_edge_attrs · Graph · Python export: yes

Attach attributes to edges from a DataFrame: each row is matched to an edge by its source and target columns and the other columns become attributes (e.g. a 'relation' label).

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).
attrs	DataFrame	One row per edge: source, target, plus attribute columns.

Outputs

Port	Type	Description
graph	networkx.Graph	A copy of the graph with the edge attributes set.

Fields

Field	Type	Default	Choices
Source column	text	source
Target column	text	target

🏷️ Set Node Attributes

id graph_set_node_attrs · Graph · Python export: yes

Attach attributes to nodes from a DataFrame: each row is matched to a node by the node-id column and its other columns become attributes (e.g. a 'type' to colour a knowledge graph by).

Inputs

Port	Type	Description
graph	networkx.Graph	The graph (from a Graph from … node).
attrs	DataFrame	One row per node: the id column plus attribute columns.

Outputs

Port	Type	Description
graph	networkx.Graph	A copy of the graph with the node attributes set.

Fields

Field	Type	Default	Choices
Node id column	text	node