Graphs and Networks: Elementary Introduction to the Wolfram Language

The result here is a list of lists. But what Graph needs is just a single list of connections. We can get that by using Flatten to “flatten” out the sublists.

Flatten “flattens out” all sublists, wherever they appear:

Flatten[{{a, b}, 1, 2, 3, {x, y, {z}}}]

Get a “flattened” list of connections from the array:

Flatten[Table[i -> j, {i, 3}, {j, 3}]]

Show the graph of these connections:

Graph[Flatten[Table[i -> j, {i, 3}, {j, 3}]], VertexLabels -> All]

Generate the completely connected graph with 6 nodes:

Graph[Flatten[Table[i -> j, {i, 6}, {j, 6}]]]

Sometimes the “direction” of a connection doesn’t matter, so we can drop the arrows.

The “undirected” version of the graph:

UndirectedGraph[Flatten[Table[i -> j, {i, 6}, {j, 6}]]]

Now let’s make a graph with random connections. Here is an example with 20 connections between randomly chosen nodes.

Make a graph with 20 connections between randomly chosen nodes numbered from 0 to 10:

Graph[Table[RandomInteger[10] -> RandomInteger[10], 20], VertexLabels -> All]

You’ll get a different graph if you generate different random numbers. Here are 6 graphs.

Six randomly generated graphs:

Table[Graph[Table[RandomInteger[10] -> RandomInteger[10], 20]], 6]

There’s lots of analysis that can be done on graphs. One example is to break a graph into “communities”—clumps of nodes that are more connected to each other than to the rest of the graph. Let’s do that for a random graph.

Make a plot that collects “communities” of nodes together:

The result is a graph with the exact same connections as the original, but with the nodes arranged to illustrate the “community structure” of the graph.

Vocabulary

Graph[{ij,...}]		a graph or network of connections
UndirectedGraph[{ij, ...}]		a graph with no directions to connections
VertexLabels		an option for what vertex labels to include (e.g. All)
FindShortestPath[graph,a,b]		find the shortest path from one node to another
CommunityGraphPlot[list]		display a graph arranged into “communities”
Flatten[list]		flatten out sublists in a list

Exercises

Check your answers in the Wolfram Cloud

21.1Make a graph consisting of a loop of 3 nodes. »

Expected output:

Out[]=

Answer & check your solution

21.2Make a graph with 4 nodes in which every node is connected. »

Expected output:

Out[]=

Answer & check your solution

21.3Make a table of undirected graphs with between 2 and 10 nodes in which every node is connected. »

Expected output:

Out[]=

Answer & check your solution

21.4Use Table and Flatten to get {1, 2, 1, 2, 1, 2}. »

Expected output:

Out[]=

Answer & check your solution

21.5Make a line plot of the result of concatenating all digits of all integers from 1 to 100 (i.e. ..., 8, 9, 1, 0, 1, 1, 1, 2, ...). »

Expected output:

Out[]=

Answer & check your solution

21.6Make a graph with 50 nodes, in which node i connects to node i+1. »

Expected output:

Out[]=

Answer & check your solution

21.7Make a graph with 4 nodes, in which each connection connects i to Max[i, j]. »

Expected output:

Out[]=

Answer & check your solution

21.8Make a graph in which each connection connects i to j-i, where i and j both range from 1 to 5. »

Expected output:

Out[]=

Answer & check your solution

21.9Generate a graph with 100 nodes, each with a connection going to one randomly chosen node. »

Sample expected output:

Out[]=

Answer & check your solution

21.10Generate a graph with 100 nodes, each connecting to two randomly chosen nodes. »

Sample expected output:

Out[]=

Answer & check your solution

21.11For the graph {12, 23, 34, 41, 31, 22}, make a grid giving the shortest paths between every pair of nodes, with the start node as row and end node as column. »

Expected output:

Out[]=

Answer & check your solution

+21.1Make a graph with 4 nodes in which every node is connected, displaying the resulting graph with radial drawing. »

Expected output:

Out[]=

Answer & check your solution

+21.2Generate a graph in which a single node is connected to 10 others. »

Expected output:

Out[]=

Answer & check your solution

+21.3Use Flatten to generate a table of the numbers 1 to 50 in which even numbers are colored red. »

Expected output:

Out[]=

Answer & check your solution

+21.4Use ImageData, Flatten and Total to find the number of white pixels in Binarize[Rasterize["W"]]. »

Expected output:

Out[]=

Answer & check your solution

+21.5Use Flatten, IntegerDigits and Total to find the sum of all digits of all whole numbers up to 1000. »

Expected output:

Out[]=

Answer & check your solution

+21.6Generate a graph with 200 connections, each between nodes with numbers randomly chosen between 0 and 100. »

Sample expected output:

Out[]=

Answer & check your solution

+21.7Generate a plot that shows communities in a random graph with nodes numbered between 0 and 100, and 200 connections. »

Sample expected output:

Out[]=

Answer & check your solution

Q&A

What’s the difference between a “graph” and a “network”?

There’s no difference. They’re just different words for the same thing, though “graph” tends to be more common in math and other formal areas, and “network” more common in more applied areas.

What are the vertices and edges of a graph?

Vertices are the points, or nodes, of a graph. Edges are the connections. Because graphs have arisen in so many different places, there are quite a few different names used for the same thing.

How is i j understood?

It’s Rule[i, j]. Rules are used in lots of places in the Wolfram Language—such as giving settings for options.

How does the Wolfram Language determine how to lay out graphs?

It uses a variety of methods to try to make them as “untangled” and “balanced” as possible. One method is essentially a mechanical simulation in which each edge is like a spring that tries to get shorter, and every node “electrically repels” others. The option GraphLayout lets you specify a layout method to use.

How big a graph can the Wolfram Language handle?

It’s mostly limited by the amount of memory in your computer. Graphs with tens or hundreds of thousands of nodes are not a problem.

Can I specify annotations to nodes and edges?

Yes. You can give Graph lists of nodes and edges that include things like Annotation[node, VertexStyle Red] or Annotation[edge, EdgeWeight 20]. You can extract these annotations using AnnotationValue. You can also give overall options to Graph. In addition, you can specify “tags” for edges using, for example, DirectedEdge[i, j, tag]. Tags are displayed if you use EdgeLabels Automatic.

Tech Notes

Graphs, like strings, images, graphics, etc., are first-class objects in the Wolfram Language.

You can enter undirected edges in a graph using <->, which displays as .
CompleteGraph[n] gives the completely connected graph with n nodes. Among other kinds of special graphs are GridGraph, TorusGraph, KaryTree, etc.
There are lots of ways to make random graphs (random connections, random numbers of connections, scale-free networks, etc.). RandomGraph[{100, 200}] makes a random graph with 100 nodes and 200 edges.
AdjacencyMatrix[graph] gives the adjacency matrix for a graph. AdjacencyGraph[matrix] constructs a graph from an adjacency matrix.
Common settings for the GraphLayout option are "SpringElectricalEmbedding" and "LayeredDigraphEmbedding".
PlanarGraph[graph] tries to lay a graph out without any edges crossing—if that’s possible.

More to Explore

Guide to Graphs and Networks in the Wolfram Language »

Next Section

More Learning

Tech Support

Educational Programs for Adults

Educational Programs for Youth

Events

Wolfram Initiatives

Educational Resources

Hobbies & Projects

Wolfram Solutions

Wolfram Solutions For Education

Get Started

Grow Your Skills

Work with Us

Read

Educational Programs for Adults

Educational Programs for Youth

Events

21 Graphs and Networks