Social network (3D)
About this app
Three classical models (Erdős-Rényi random graph, Watts-Strogatz small-world, Barabási-Albert scale-free) live in a 3D force layout. Clicking a node highlights its neighbours; the statistics panel shows degree, clustering and path lengths live.
Three classic network models
Network science describes social relationships as graphs of nodes and edges. Three models are classics. In the Erdős–Rényi random graph each pair of nodes is linked independently with probability p, which gives short paths but little clustering. The Watts–Strogatz model starts from a ring of local links and rewires edges with probability p; a handful of shortcuts makes paths much shorter while clustering stays high. That is the small-world effect behind “six degrees of separation”. In the Barabási–Albert model new nodes prefer to attach to well-connected ones (preferential attachment), producing a few hubs with very high degree.
What the app shows
The app generates networks of 20 to 150 nodes and lays them out live in a 3D force-directed layout. Each model has one parameter: the edge probability p for the random graph, the rewiring probability for the small-world model (each node starts with four ring neighbours) and the number m of new edges per node for the scale-free model. A statistics panel shows nodes, edges, mean and maximum degree, clustering coefficient C, mean path length and number of components. The network can also be built up step by step.
Try it yourself
Start the small-world model with a rewiring probability of 0: clustering is high and paths are long. Raise the value only slightly and watch the mean path length drop sharply while clustering stays high at first. In the random graph, a very small p leaves many separate components that merge as p grows. In the scale-free model, keep an eye on the maximum degree. Click a node to highlight its neighbours and compare a hub with a peripheral node.
Frequently asked questions
- What is a small-world network?
- A small-world network combines high clustering, meaning many triangles among acquaintances, with short average paths between any two nodes. Watts and Strogatz showed in 1998 that a few random shortcuts in an otherwise locally connected graph are enough. Many social networks have this property.
- What does scale-free mean for a network?
- In a scale-free network the degree distribution roughly follows a power law: most nodes have few links, a few hubs have very many. The Barabási–Albert model produces such networks through growth and preferential attachment to well-connected nodes. Web links and citation networks are common examples.
- What does the clustering coefficient measure?
- The clustering coefficient measures how likely two neighbours of a node are to be connected to each other, in everyday terms whether your friends are also friends with each other. It ranges from 0 to 1. Random graphs usually have low clustering, social networks much higher.
- What can I do in this network app?
- You choose between the random graph, small-world and scale-free models, set the number of nodes and the model parameter, and see the network in 3D. Metrics such as degree, clustering and path length update live. Clicking a node highlights its neighbours.
Subject: Sociology & networks
More from Sociology & networks
A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps