# Industry Connections Network

<details>

<summary>Definition</summary>

The Industry Connections Network illustrates how sub-industries are interrelated through shared capabilities, co-location, and sectoral overlap. It captures the territorial story of an economic system by revealing how local industries interact, cluster, and integrate. This network visualizes aggregated sub-industries, each mapped to a 6-digit NAICS code. This level of aggregation allows for a strategic view of how economic activity is structured across cities in the metro region. Industries with many connections tend to anchor supply chains or enable cross-sector innovation, while more isolated industries may offer specialized or emerging capabilities with limited integration.

</details>

<details>

<summary>Navigation</summary>

This tool is presented as an interactive Force-Directed Diagram (FDD). Each metro area city has its own standalone network, switchable via a dropdown or node interaction.

* Nodes represent sub-industries (mapped to [NAICS 6-digit codes](https://www.naics.com/six-digit-naics/)), color-coded by NAICS 4-digit industry classification, sized by revenue.
* Links indicate relationships between businesses, based on shared characteristics and co-occurrence
* Clicking on a node highlights its most direct connections

</details>

<details>

<summary>Methodology</summary>

To build the Industry Connections Network, businesses are mapped into a sectoral graph using the following steps:

1. Business Aggregation - All businesses are geolocated and assigned a 6-digit NAICS code. These businesses are then aggregated into sub-industries per city or metro area.
2. Node Creation - Each unique NAICS 6-digit code becomes a node in the network.
3. Edge Construction - Connections between nodes are determined using co-occurrence patterns. If two industries frequently appear together in the same city or region across the dataset, they are considered “close” A similarity threshold is applied to determine which pairs are strongly connected enough to form an edge
4. Graph Layout - A force-directed layout is used, which simulates physics: more tightly connected industries are drawn closer together, while loosely related ones drift apart. This spatialization reveals natural clusters, central sectors, and structural g

</details>

<details>

<summary>Calculation</summary>

This tool is descriptive rather than numerical, but underlying it are network science principles like:

* Degree Centrality – Number of connections a firm has (how integrated it is within its local economic network)
* Modularity / Community Detection – Grouping firms into distinct economic clusters
* Connected Components – Isolated vs. integrated firm groups
* Force-directed algorithms – Visual logic used to spatially represent proximity based on relationship strength

There’s no single KPI value per unit here, but the structure helps reveal qualitative network position.

</details>

<details>

<summary>Interpretation</summary>

* Central Nodes represent firms with many connections, often key players or anchors in local supply chains.
* Peripheral Nodes show smaller or more specialized firms with fewer visible ties that may still provide unique value but are less integrated.
* Dense Clusters suggest strong local specialization or co-located ecosystems (e.g., a biotech corridor or advanced manufacturing hub).
* Sparse or fragmented areas may indicate economic isolation, weak integration, or underutilized industrial capacity.

The network helps:

* Spot strategic sectors and business clusters
* Identify opportunities for collaboration or value chain development
* Understand which firms serve as connectors in the broader economy
* Inform economic diversification and industrial resilience strategies

</details>