1982
Billie Jean
Michael Jackson
▶ Click to play1980 – 1999
In the album and radio era, popular music lived inside slower systems: physical releases, radio programming, album sequencing, and fewer instant exits. That environment did not force every song to be long, but it gave many hits more permission to build gradually.
2000 – 2014
Downloads, iPods, and early streaming changed the unit of listening. Songs became easier to search, save, shuffle, and skip. The album still mattered, but the individual track was becoming the main object moving through the music economy.
2015 – Present
If Spotify becomes more personalized around 2015, do popular songs start behaving differently after that point? The useful question is whether shorter hit durations and explicit-content changes line up with the platform shift.
This project treats 2015 as context, not proof of causation.
1982
Michael Jackson
▶ Click to play2024
Lady Gaga & Bruno Mars
▶ Click to playThese two songs are not proof by themselves. They make the larger question easier to hear.
Songs are getting shorter.
Explicit hits are getting more common.
Scroll to see the evidence.
Random hit track
Choose a year or press random.
Audio panel: random hit song from the current scope, embedded from Spotify.
Chart 1: Average track duration (minutes) for yearly hits. Click a point to filter Chart 2.
Chart 2: Distribution of track duration (minutes) for all songs.
Use the tabs to inspect one musical feature at a time. The chart compares yearly hits against the broader catalog, using the same hit threshold as the duration graphs above.
Feature explorer: choose a tab to compare how hit songs and the wider catalog changed over time.
Search for a song to see its duration, explicit flag, audio features, and chart position.
We assembled and cleaned Spotify track data spanning decades, then built the page as a chapter-based explorable explanation rather than a standalone dashboard. The prototype now begins with two listening examples, moves through a short history of how music discovery changed, and then uses charts to test whether the story appears in the data. Users can adjust the hit threshold to test how strict or broad the definition of a "hit" should be, click a year to inspect its distribution, and compare the duration trend against explicit-content changes.
The most challenging design problem is balancing narrative clarity with evidence. The 2015 Spotify milestone gives the project a compelling turning point, but it should be framed as context rather than proof of direct causation. Features like duration, explicitness, energy, danceability, release strategy, and popularity each capture different dynamics, so the final project needs to guide users through them without implying that one metric explains everything.
Note: This interaction uses observed dataset values only. "Hit" is defined operationally by your chosen Spotify popularity percentile in each year.