Overview

The Web of Trust Map is a public reference for the decentralised digital identity ecosystem. It connects around 9,500 projects, consortia, blockchains, public and private organisations, people, regulations, standards, DLT technologies and DID methods in an interactive graph.

Visitors can explore the network, filter it by geography, industry and entity type, and open a profile for each item. The same underlying research is available through a public REST API and as an openly licensed dataset, allowing journalists, policymakers and researchers to use the data outside the visualisation.

The site is entirely public and requires no account. Alongside the graph, it provides guided explanations and aggregate views that make a specialised technical field legible to people who do not already know its terminology.

My work

I built the Web of Trust Map for Key State Capital from its data model and application through graph rendering and production infrastructure, working with a designer on the visual design.

Turning research into structured data

The starting point was an inherited backend setup: research was collected in SharePoint and copied into Postgres through Fivetran. There was no application, and the imported schema was severely denormalised and incomplete. Before building the visualisation, I had to model ten interrelated kinds of real-world entities, retain their sources and resolve when differently named records describe the same thing.

I replaced the inherited pipeline by moving collection into the primary Postgres database and removing Fivetran. The resulting data model powers the map, public API and dataset, while internal tools and AI-assisted pipelines support classification, description and human review.

Making the graph responsive

Drawing thousands of interconnected entities is substantially more expensive than displaying their profiles or searching the dataset. The graph must also remain usable on older hardware. The guiding approach is to precompute expensive work at every layer so the visitor's browser can concentrate on drawing.

Large layouts are calculated ahead of time by a headless browser driving the real application through relevant filter combinations. The resulting node positions are cached and sent to visitors as coordinates, avoiding a full force simulation on each device. Smaller views remain dynamic, where the movement is useful and browsers can calculate it without difficulty.

Expensive relationship aggregates are prepared in Postgres, while the graph payload and query results are cached before reaching the browser. Rendering then adapts to the current device and graph size: slower devices render at a lower pixel density, and decorative interactions are progressively reduced as more nodes become visible. This keeps the complete dataset navigable without replacing it with a simplified sample.

Today

The Web of Trust Map publishes around 9,500 entities across ten types, a public API and the complete dataset under a CC BY 4.0 licence. The project is live for Key State Capital, with its launch distributed through GlobeNewswire, Barchart and openPR.