Calculator
Enter nodes and edges, then choose graph rules.
Formula used
Network density measures how close a graph is to being complete.
- Directed (no self-loops):
D = E / (N(N−1)) - Directed (with self-loops):
D = E / (N²) - Undirected (no self-loops):
D = 2E / (N(N−1)) - Undirected (with self-loops):
D = 2E / (N(N+1))
This calculator also reports maximum possible edges and average degree using 2E/N (and E/N for directed in/out averages).
How to use this calculator
- Enter the total number of nodes N.
- Enter the total number of edges E.
- Select Directed or Undirected to match your network.
- Choose whether self-loops are permitted in your model.
- Click Calculate to see results above the form.
- Use Download CSV or Download PDF after calculation.
Example data table
| Case | N | E | Graph type | Self-loops | Max edges | Density |
|---|---|---|---|---|---|---|
| A | 10 | 18 | Undirected | No | 45 | 0.400000 |
| B | 10 | 18 | Directed | No | 90 | 0.200000 |
| C | 5 | 8 | Undirected | Yes | 15 | 0.533333 |
Examples assume a simple count of edges under the stated rules.
Network density in practice
Network density summarizes how many connections exist compared with the maximum possible under your graph rules. It is dimensionless, ranges from 0 to 1, and helps compare networks of different sizes in a consistent way.
1) What density means
A density near 0 indicates a sparse network with few observed links. A density near 1 indicates a highly saturated network where most allowable links are present. Because density is normalized by the maximum edge count, it is more comparable than raw edge totals.
2) Maximum possible edges
The upper bound depends on whether your network is directed and whether self-loops are permitted. For example, an undirected simple graph allows N(N−1)/2 edges, while a directed graph without loops allows N(N−1) ordered links. This calculator reports that maximum explicitly.
3) Scaling with network size
As N grows, the number of possible edges increases on the order of N². That growth can make large networks appear sparse even when the edge count is large. Density corrects for this by dividing by the theoretical maximum, letting you compare connectivity across different system sizes.
4) Directed vs undirected interpretation
In an undirected network, each edge contributes to two node degrees, so average total degree is 2E/N. In a directed network, edges split into in-links and out-links, with averages E/N each. Density is still computed as observed links divided by the directed maximum.
5) Self-loops and modeling choices
Allowing self-loops increases the maximum edge count and can reduce density for the same E. In physical interaction networks, loops might represent self-interaction, feedback, or reflection paths. If your data does not contain such links, keep loops disabled to avoid understating connectivity.
6) Using missing edges as a diagnostic
Missing edges equals maxE − E and indicates how far the network is from saturation. In experiments, a large missing-edge value can signal limited coupling, incomplete sampling, or strong constraints that prevent many interactions. Track this value when testing how connectivity changes over time.
7) A numeric example
Suppose N = 50 and E = 300 for an undirected simple network. The maximum is 50×49/2 = 1225, giving density 300/1225 ≈ 0.2449. If the same counts were directed without loops, the maximum would be 50×49 = 2450, and density would be about 0.1224.
8) Reporting and comparison tips
When comparing systems, report density together with the rule set (directed or undirected, loops allowed or not) and the node count. For noisy data, repeat measurements and summarize density with averages and ranges. Export CSV for logs and PDF for quick reporting.
FAQs
1) Can density be greater than 1?
No. Density is normalized by the maximum possible edges, so it should be within 0 to 1. If your edge count exceeds the theoretical maximum for your selected rules, the calculator flags it as a check.
2) Why does density drop when I enable self-loops?
Allowing self-loops increases the maximum possible edges. If E stays the same, the ratio E/maxE becomes smaller, so the reported density can decrease.
3) What should I use for multigraphs with parallel edges?
This calculator assumes a simple edge count under the chosen rules. If your data has parallel edges, consider collapsing them into a single connection per node pair, or treat the network as weighted and analyze strength separately.
4) How does density relate to average degree?
Average degree scales with E/N, while density scales with E/N² through the maximum edge term. Two networks can share a similar average degree but have different densities if their sizes differ.
5) Is density meaningful for very small networks?
Yes, but it changes in coarse steps when N is small. For tiny graphs, also report E and maxE so readers can see the discrete structure behind the ratio.
6) What does a very low density imply physically?
It often indicates weak coupling, limited interaction opportunities, or incomplete observation. In many physical and biological systems, sparse connectivity is expected because constraints restrict which nodes can interact.
7) Which export should I choose?
Use CSV when you want to store results, batch compare cases, or import into analysis tools. Use PDF when you need a clean snapshot for sharing, lab notes, or quick reporting.