Skip to main content

INTELLIGENCE FUSION · NEXUS

Reveal the hidden network — the broker you didn't know to look for

Nexus is the i2/Gotham-class link chart, with one decisive difference: it doesn't wait for someone to hand it a structured dataset. Sentinel derives the relationship graph automatically from what your cameras already saw — then the central figure who bridges two otherwise-separate cells surfaces on its own.

Nexus · Link Analysis
2 cells · 1 broker
BROKER
Seed Broker (bridge) MemberSeed two names — the hidden broker surfaces itself.

Seed two peripheral names — the bridge between the cells surfaces automatically.

Auto
Derived from footage
The graph builds itself from camera co-occurrence — no structured dataset to feed in
1 broker
The bridge you missed
Betweenness surfaces the single person who links two otherwise-separate cells
1–3 hops
Ego expansion
Seed N subjects, expand over the connection backbone, edges banded by tie strength
Reproducible
Built for review
Deterministic metrics return the same graph on every run

THE DIFFERENTIATOR

Most platforms need the network handed to them. Sentinel finds it.

A conventional link-analysis tool is only as good as the structured intelligence someone loads into it — a spreadsheet of known associates, a case file, a prior investigation. Sentinel starts a step earlier. It reads co-presence straight off the camera estate and turns it into a weighted relationship graph, so the network exists the moment your cameras do. That is the video-native wedge — you don't need to already know who is connected to build the chart that shows you.

Built from what the cameras saw

Every time two people appear at the same camera in the same window, that co-presence becomes a weighted tie. The relationship graph is a by-product of watching — not a dataset you have to assemble and import first.

No pre-built dataset required

You do not have to already know the network to reveal it. Nexus derives structure from raw sightings, so the first useful chart appears before any analyst has typed a single known association.

Deterministic and reproducible

The centrality, bridge and community metrics are computed deterministically — the same seeds return the same graph every time. Reproducibility is a forensic requirement, and it is built in, on your own infrastructure.

THE OUTCOME

Seed two peripheral names. The broker surfaces itself.

This is the move a plain video system can never make. Drop in two subjects you already care about — even two people who never appear on camera together — and Nexus expands the network around each of them across the connection backbone. Where the two neighbourhoods meet, the person who quietly bridges them lights up as a broker. It is the target you didn't know to look for.

  • Betweenness centrality identifies the node that the most paths route through — the bridge between two otherwise-separate cells — and highlights it automatically.
  • Nodes are coloured by role — seed, broker, cut-out and community — so the shape of the network reads at a glance, not after hours of manual layout.
  • Edges are banded by tie strength, so a strong, recurring relationship looks different from a one-off crossing.
  • A node inspector shows degree, betweenness and role, opens the full profile, and lets you add any node as a new seed to pivot the investigation.
Nexus link-analysis board with seeded subjects, the broker node highlighted, tie-strength edges and a node inspectorClick to enlarge

Seed two peripheral names — the broker who bridges the two cells surfaces on its own.

WHAT NEXUS SURFACES

The five questions a link chart should answer

Nexus is more than a pretty force-layout. Behind the board sits a graph-traversal engine that answers the questions an investigator actually asks of a network — each computed over a materialised edge backbone, not a slow per-query scan, so the chart stays responsive at national scale.

BROKER / BRIDGE

Who holds the network together

Betweenness centrality finds the broker — the figure the most connections pass through. Remove them in your mind and the network loses its glue. It is usually the person no single watchlist would have flagged.

CUT-OUT

Whose removal breaks the cell

Articulation-point analysis finds the cut-out: the single node whose removal fragments a connected cell into separate pieces. It is the structural weak point — and the highest-value place to focus.

COMMUNITIES

The clusters inside a crowd

Connected-component and community detection separate one dense mass of sightings into the distinct groups that actually move together, so a single event is not mistaken for a single network.

STRONGEST PATH

How two people are really linked

Pick person A and person B and Nexus finds the strongest chain between them — favouring a run of strong ties over one long, weak hop — with per-hop evidence you can open.

TIE STRENGTH

A connection crowd noise cannot fake

Sharing a busy thoroughfare does not make two people connected. Sentinel weighs each tie so that repeated, off-peak, moving-together encounters count and incidental crossings at a hub do not — real links stand out from the crowd.

PIVOT

Every node is a new starting point

Add any surfaced node as a seed, tighten the minimum tie strength, or step the depth from one to three hops — the board re-expands live so the investigation follows the evidence wherever it leads.

STRUCTURAL WEAKNESS

Find the cut-out that holds a cell together

Some networks survive losing almost anyone. Others hinge on a single person. The cut-out is the node whose removal splits a connected cell into fragments — and Nexus finds it deterministically, without an analyst tracing edges by hand. Knowing where a network is fragile is knowing where to concentrate.

  • Articulation-point analysis marks the nodes that are load-bearing for the whole structure, not just well-connected.
  • The board visibly shows what the network becomes without that node — one cell splitting into two — so the finding is self-evident, not a number to trust on faith.
  • Community colouring stays consistent as you expand, so a cluster you identified at depth one is still recognisable at depth three.
  • Because the metrics are deterministic, the same finding reproduces for a colleague, a supervisor, or a court — every run returns the same graph.
Nexus board expanded to three hops — communities coloured, the load-bearing cut-out visible where the cell would fragmentClick to enlarge

Expand the graph and the load-bearing node — the one whose removal splits the cell — reads at a glance.

HOW IT COMES TOGETHER

From raw sightings to a network that explains itself

No manual charting, no dataset to import, no analyst placing nodes by hand. The relationship graph is a standing by-product of the camera estate — Nexus just reads it back to you with the important roles already coloured in.

1

The cameras do the linking

Co-presence at a camera in a shared time window becomes a weighted tie between two people. The graph accumulates continuously from ordinary footage — nothing has to be entered by hand.

2

Crowd noise is filtered out

A busy thoroughfare cannot fake a connection. The weighting discounts hub cameras and rewards genuine, repeated, off-peak, moving-together encounters, so real relationships stand above the background.

3

Seed the people you know

Drop in two or more subjects and Nexus expands the network around them across the connection backbone, one to three hops out, banding every edge by tie strength.

4

The hidden roles light up

Brokers, cut-outs and communities are computed and coloured automatically — including the bridging figure you did not know to look for. The chart explains the network instead of just drawing it.

See the network in your own footage

Nexus turns the cameras you already run into a relationship graph that names the broker, the cut-out and the communities — sovereign, on your infrastructure, reproducible for review. Book a demo and watch a hidden bridge surface itself.