Poster Index

Industry

Data Visualization

Year

2026

An interactive 3D visualization of posters from the Victoria and Albert Museum's open API. (Work in progress)

Services

D3.js
p5.js / ml5.js
Three.js
Figma

Brief

Users can explore over 16,000 artworks by artist, place of origin, category, and dominant color. The project was originally built with p5.js and the ml5.js machine learning library to classify poster images from the V&A's open access API, then brought into an interactive 3D space with Three.js, where posters cluster by classification at adjustable confidence levels.

Finding

A dark theme lets the posters take visual priority, and an introductory animation grounds users in the collection's metadata before they navigate the 3D space. The clustering got stronger once it responded to user-driven filters rather than fixed categories — that shift is what's currently being refined, along with integrating Vibrant.js to extract each poster's dominant color.

Work in Progress

Still refining the clustering — and that's the point.

The project is an ongoing exploration: the 3D clustering is being reworked around categorical metadata rather than fixed groupings, and color extraction via Vibrant.js is still being integrated. The metadata shown in the intro animation is currently placeholder and doesn't reflect the final dataset yet.

Full Project Walkthrough

Clicking into an individual poster.

Selecting a poster from the 3D space pulls its V&A record forward — title, origin, and category metadata alongside the dominant colors driving its placement in the cluster.

Next Project Breathscape