Web Application / 2026

attention field

Product design · Interaction · Engineering

A live Wikipedia edit stream translated into motion, weight, and proximity, so emerging patterns can be felt at a glance instead of disappearing into a list.

What if live data could be felt and understood at a glance, instead of read one update at a time?

A conventional feed gives every update its own line and asks the viewer to assemble the larger picture. Attention Field approaches that problem spatially: busy articles carry more visual weight, related topics gather, and movement suggests shifting activity. The goal is an immediate, intuitive impression of the data, something you can sense before deciding what deserves a closer look.

A live editorial-activity instrument built from English Wikipedia's public recent-change stream. It is an exploration of Wikipedia editing behavior, not a claim to measure public opinion.

01

Follow the signal

Stream recent edits through Wikimedia EventStreams, with activity, topic classification, and edit history driving the display.

02

Make relationships perceptible

Use real shared Wikipedia categories to draw articles together with springs, repulsion, and collision. Proximity and motion make clusters readable without asking someone to compare rows in a table.

03

Interrogate the patterns

Analyze edit summaries with a lightweight pattern engine that looks for repeated language across distinct editors, and expose a second, index-based view of the same data.

Intuition must mean something

Activity determines size, topics determine color, and genuine connections shape movement. An attractive but ungrounded glow was removed: visual immediacy is only useful if it expresses the underlying data.

Verification without guesswork

Standalone tests exercised the simulation against real data, catching physics errors before release; the topic palette was checked for accessibility.

A live browser instrument that offers two complementary encounters with the same activity: FIELD for sensing emerging patterns quickly, and INDEX for inspecting the individual records behind them. It is a design experiment in making complex information approachable before making it analytical.

What this does not claim

An expressive view of a particular public editing stream, not a statistically representative account of what everyone is paying attention to.