American Emotions

Public attention, rendered as color.

American Emotions. Public attention interpreted as a luminous color field.
Public attention, interpreted through color.

Attention leaves
a trace.

The project began during the 2024 election period. Public search activity supplied a signal; an authored model assigned emotional weights and colors. Each interpretation became part of a field.

An archival rendering from the 2024 election-period experiment.
The archival plate: American Emotions, 185 days.

A record,
over time.

The individual traces accumulate. Read together, they make a visual record of the model’s interpretations of public attention.

The field carries those interpretations. It does not establish how a population felt.

Project details

Origin
2024
Scope
United States
Form
Instrument
Continuation
Global Emotions
About the project

An authored instrument that interprets public attention through emotion and color.

The project began in 2024 with a question: could public attention be interpreted as an affective field, and what would the right medium be? Color became a way to explore that question.

The original election-period experiment used search activity as a proxy for attention. Its fields record the computational model’s interpretation of those signals, rather than a verified measurement of a population’s emotional state.

Explore the successor: Global Emotions
Signal → Interpretation → Rendering

Signal

  1. Public search activity
  2. Google Trends RSS

Interpretation

  1. Authored affective model
  2. Emotion weights and color assignments

Rendering

  1. Accumulation over time
  2. A luminous color field

Each query receives emotional weight and a color within the authored system. Its trace joins a field that can be read as an atmosphere. Search activity supplies the input; the model supplies the interpretation.

Rendering method

The later American Emotions renderer uses float32 additive accumulation with Gaussian bloom. Individual queries leave soft traces that layer into a density map of interpreted attention.

Taxonomy across versions

The retained 2024 pipeline asks a language model to assign emotion labels and RGB colors using an open vocabulary. The later American Emotions v2 source defines 171 emotions; Global Emotions documents 169. These describe different iterations, not interchangeable counts for the original experiment.

What the experiment opened

The 2024 experiment opened questions about the distance between a public signal, an emotional interpretation, and a visual form. What can attention stand in for? What does the model introduce? What can color communicate that a label cannot?

That line of inquiry later informed Emotion as System, Chroma, and the ACG installation series. Chroma gives the person authority to name a feeling; the installation places participant input beside interpreted public signals. Each returns the original question in a different form.

New York EmotionsInferred emotion meets feelings expressed in personEmotion as SystemThe research foundationChromaA private emotional journalTell Me How You Feel: ACGA shared encounter with affect
From American Emotions to Global Emotions

Global Emotions continues this inquiry across locations in a daily field interface. The instrument is a later development, not a live replay of the 2024 American Emotions system. Its model and coverage should be read through its own dated records.

Global EmotionsPublic signals interpreted across locations
Project record
Origin
American Emotions · 2024 election-period experiment
Related work
New York Emotions · 2025 mayoral-election period
Continuation
Global Emotions · Location-based instrument
Research & connections
All work →