Social Network Centrality Calculator

What Is the Social Network Centrality Calculator?

The Social Network Centrality Calculator computes degree, betweenness, and eigenvector centrality together from a simple list of connections, the three measures researchers most commonly use to identify which nodes in a social or organizational network and the community size and stability calculator matter structurally, and why. Type in pairs of connected people, accounts, or entities, one per line, and the calculator builds the network automatically and ranks every node on all three measures at once. Existing free tools for this calculation are either full graph-analysis software such as Gephi and NodeXL, which require installation and a learning curve, or programming libraries such as NetworkX, which require writing code; no simple browser-based calculator computing all three centrality measures from a typed edge list existed publicly at the time this tool was built.

Centrality measures are used across sociology, organizational research, public health contact tracing, and information-spread modeling, given that the same underlying question, which nodes structurally matter most, applies whether the network represents friendships, company communication patterns, or disease transmission paths. Because no single centrality measure captures every notion of importance, comparing all three at once is standard analytical practice rather than a shortcut.

Degree, Betweenness, and Eigenvector: Three Different Definitions of Important

Degree centrality counts a node's direct connections and nothing else, making it the simplest and most local measure: a node with five connections always outranks a node with three, regardless of who those connections are. Betweenness centrality instead counts how often a node sits on the shortest path between two other nodes, identifying bridges and bottlenecks that control the flow of information or resources through the network, calculated here using Brandes' efficient shortest-path algorithm rather than the slower brute-force approach of enumerating every path directly. Eigenvector centrality goes a step further than degree by weighting each connection by how centrally positioned that neighbor is, calculated through power iteration on the network's adjacency matrix, so a node connected to a few highly influential neighbors can outscore a node with many low-influence ones.

Why a Node Can Rank High on One Measure and Low on Another

The clearest illustration is a bridge node connecting two otherwise separate clusters: it may have only two or three direct connections, giving it a low degree centrality score similar to any peripheral node, while sitting on nearly every shortest path between the two clusters, giving it the highest betweenness centrality in the entire network. This calculator flags exactly that pattern automatically whenever the top-ranked node differs between degree and betweenness. A separate, equally common pattern shows up with eigenvector centrality: a node connected to many disconnected, low-influence neighbors can have a respectable degree score but a low eigenvector score, since its neighbors are themselves poorly positioned in the network.

The calculator's own default example demonstrates this directly: seven nodes form two triangles, A-B-C and D-E-F, joined only by a single connector node, Bridge, linking C to D. Counting direct connections, Bridge has only two, identical to several peripheral nodes in either triangle, so a degree-only ranking buries it in the middle of the list. Betweenness tells a different story: every shortest path from any node in the A-B-C triangle to any node in the D-E-F triangle must pass through Bridge, since no other route exists between the two halves of the network, giving Bridge the highest betweenness score by a wide margin despite its low degree. Removing Bridge from this network splits it into two completely disconnected triangles, the clearest possible illustration of why betweenness, not degree, identifies which node a network actually depends on structurally.

From Theory to Real Networks: Resilience and Removal

Network resilience research consistently uses betweenness centrality to identify which single node's removal would most damage a network's connectivity, since removing the highest-betweenness node, particularly a sole bridge between two clusters, can split a connected network into separate disconnected components or dramatically lengthen the paths between them. This is directly useful for identifying single points of failure in organizational communication structures, supply chains, or any network where you want to know not just who is well-connected, but whose absence would be structurally costly. For populations where you also want to measure overall social cohesion rather than individual node importance, our Relative Deprivation Index Calculator approaches network-adjacent inequality from the income-comparison side of the same broader sociological toolkit.

Accuracy and Limitations

All three measures here are calculated exactly for the network you enter, but centrality is a measure of structural position, not a guarantee of real-world importance. A 2019 study published in Scientific Reports found that the correlation between most centrality measures and actual causal influence in a system is weak, with eigenvector centrality showing the strongest, though still imperfect, relationship among the standard measures. Eigenvector centrality specifically should be interpreted with care on networks containing multiple disconnected components, since the power iteration method used to calculate it concentrates scores within the largest connected component and can understate the importance of nodes in smaller, separate clusters.

The Most Common Centrality Mistake

The mistake I see most often is treating degree centrality, the easiest measure to compute by hand from a simple connection count, as if it were a complete answer to "who matters most" in a network. Counting direct connections alone consistently misses bridge nodes, the people or accounts with relatively few direct ties who nonetheless connect otherwise separate groups, and whose removal can fragment a network far more severely than removing a highly-connected but centrally embedded node would. Whenever someone hands me a network analysis based on a simple connection-count ranking alone, I check the betweenness ranking before accepting any conclusion about which nodes are structurally critical, since the two rankings frequently disagree on exactly the nodes that matter most for resilience.