MAP A
ONE AUDIO · TWO SEMANTIC MAPS · OBSERVED OVER TIME
Watch a sound find
new cultural homes.
We track how the hashtag and topic context surrounding the same TikTok audio changes across a fixed observation window.
THE DUAL ATLAS
The same videos, seen through two kinds of meaning.
Audio file not addedDataset video ↗
OBSERVED USES—
HASHTAG FAMILIES REACHED—
TOPIC FAMILIES REACHED—
MAP B
Topic-family UMAP
Each point is one observed video.Nearby points use similar language.Axes have no direct meaning.Hover or click a highlighted point to find it in both maps.
USAGE OVER TIME
Videos using this audio each day
—DRAG TO SCRUB OBSERVED DAYS—
METHOD & BOUNDARIES
Two fixed landscapes.
One synchronized clock.
Both UMAPs are calculated across the shared corpus. Selecting an audio or moving through time changes which observed videos are emphasized; it does not move the underlying coordinates.
Hashtag families come from hashtag language. Topic families come from caption context. Proximity represents semantic similarity inside each map, not influence or transmission.
REFERENCES
Data, methods, and memetics.
- 01
TikTok-10M dataset ↗
The public source dataset associated with the analyzed records.
- 02
McInnes, Healy & Melville · UMAP ↗
The dimensionality-reduction method used for both semantic maps.
- 03
Pilipets & Chao · Noise in sonic social media ↗
Sound-linking, hashtagging, and memetic soundscapes on TikTok.
- 04
Oliveira et al. · Epidemic wave dynamics of music virality ↗
A temporal account of music virality across online platforms, including TikTok.
- 05
Goel et al. · The structural virality of online diffusion ↗
A framework distinguishing broadcast reach from multi-generation diffusion.
- 06
Calcifer Computing · Memetics track ↗
The track framing and interaction reference for this project.