🧠5D Political Compass
I asked Claude to create a 3D political compass visualization using Swedish political parties and their Wikipedia pages, and the result turned out pretty interesting. You can check it out here.
The Experiment
The idea was simple: take all the political parties registered for the Swedish 2026 election, analyze their positions using AI, and plot them in three-dimensional space instead of the traditional two-dimensional political compass.
Beyond Left vs Right
Most political compasses only show economic (left-right) and social (authoritarian-libertarian) axes. This version adds a cultural axis spanning from traditional to progressive values - capturing positions on issues like LGBTQ+ rights, immigration, multiculturalism, and gender roles that don’t fit neatly into the traditional categories.
Comprehensive Coverage
The visualization includes every party registered for the Swedish 2026 election - not just the major ones. Parties currently in the Riksdag are highlighted with full opacity, while smaller parties appear slightly faded but are still fully interactive. Each party links to both their homepage and Wikipedia page.
AI-Generated Positioning
Claude analyzed each party’s Wikipedia page and generated their positions on all three axes, providing detailed explanations for each coordinate. You can see these explanations in the popup when you click on any party.
Methodology Notes
The positioning is based on AI analysis of Wikipedia content, which introduces potential biases from both the AI model and Wikipedia’s editorial tendencies. The transparency of this process is important for interpreting the results.
Technical Implementation
The whole project took about a day of collaborative coding with Cursor and Claude. The 5D compass is rendered as a 3D visualization using Three.js with interactive hover effects, smooth camera controls, and dynamic label positioning.
Results
The visualization provides an interesting perspective on how Swedish political parties distribute across ideological space when viewed through this expanded dimensional framework. Some clustering patterns and outliers become apparent that might not be obvious in traditional left-right analysis.