A Multi-Level Visual Analytics Approach to Artist-Era Alignment in Popular Music
This work addresses the need for more nuanced artist-level analysis in computational music studies, though it is incremental as it builds on existing methods for era-specific baselines.
The paper tackles the problem of interpreting artist-level stylistic alignment with historical eras in popular music, introducing a visual analytics framework that reveals how alignment and intensity diverge across artist trajectories.
Existing computational studies of popular music primarily model aggregate trends or predict chart performance, offering limited support for interpreting artist-level alignment against historical stylistic baselines. We introduce an interactive visual analytics framework that treats each artist-decade as a unit defined relative to an era-specific baseline, characterized along two complementary dimensions: profile shape similarity, capturing directional correspondence with the era's feature pattern, and profile contrast ratio, capturing stylistic intensity relative to the era's dispersion. Together, these dimensions define a quadrant-based trajectory space for reasoning about conformity, divergence, and amplification over time. Applied to weekly U.S. Billboard Hot 100 chart entries from the all-time top-10 artists across six decades (1960s-2010s), linked with Spotify audio features, the framework reveals that alignment and intensity can meaningfully diverge across artist trajectories.