Building the AI Influence Map

The workflow and tech behind my data visualization project for Bloomberg Government.

This project began with an ambitious goal: to build a comprehensive database of the people, organizations, financial flows, and influence networks shaping AI policy, and to visualize the connections between them.

Months of research, analysis, and data collection from a mix of public and proprietary data sources culminated in our relational database. From there, I started designing the network visualization, building high-fidelity mockups and prototypes in Figma. I explored JS libraries built specifically for network visualizations like Cytoscape.js, but settled on D3.js for the control it gives you at a low level.

Tech stack

  • Vue & Vite: frontend
  • D3: force-directed graph
  • Postgres: database
  • Node.js: backend

The challenge

The project needed to plot a complex web of connections between hundreds of people and organizations without feeling visually overwhelming. I focused on:

  • Visual clarity: Highlighting the most connected nodes to guide users and make interactive elements immediately obvious
  • Ease of use: Implementing a Voronoi overlay around each node to create a larger hoverable and clickable target area
  • Responsive design: Building an experience that feels robust and intuitive across devices and screen sizes
  • Performance: Rendering the graph on Canvas rather than SVG to handle hundreds of nodes efficiently and ensuring D3 only emits tick events that are absolutely necessary

What’s next?

The AI Influence Map is an ongoing project. Version 1.0 focuses on 25 of the largest companies spanning the full AI ecosystem, but the next goal is to expand the visualization and its universe of connections.