Doldur Music interactive story
Music streaming habits 2026.
What can the data prove?
This Music Streaming Habits 2026 analysis lets you filter 4,000 made-up listener profiles across platforms, genres, moods, and subscriptions. It also tests a prediction model and listener clusters without pretending synthetic data describes real people.
View the Kaggle sourceBuild your view
Pick a group.
See what moves.
Choose a country, age, or subscription to update the story. The plan comparison keeps every option visible and highlights your choice. These fictional profiles are for exploration—not conclusions about real listeners.
All 4,000 fictional profiles are in view.
01 · Platforms
Where this sample
listens.
The tallest bar changes with your filters. This is the mix of fictional profiles on screen, not real-world market share. For measured listener research, compare the samples in our 2026 music streaming statistics guide.
Platforms in view
4,000 fictional listeners in view
See platform numbers
| Platform | Listeners | Share |
|---|---|---|
| Spotify | 1,591 | 39.8% |
| Apple Music | 906 | 22.7% |
| YouTube Music | 689 | 17.2% |
| Amazon Music | 315 | 7.9% |
| SoundCloud | 278 | 7.0% |
| Tidal | 221 | 5.5% |
02 · Taste
What rises
to the top.
Change the filters to see which genres, artists, and moods move up or down inside this fictional world. The rankings are a demonstration, not a map of regional taste. Explore ways to discover new music and what research says about music algorithm filter bubbles.
Genres at the top
4,000 fictional listeners in view · Every country
See genre numbers
| Genre | Listeners | Share |
|---|---|---|
| Pop | 721 | 18.0% |
| Hip-Hop | 605 | 15.1% |
| Rock | 472 | 11.8% |
| EDM | 382 | 9.6% |
| R&B | 341 | 8.5% |
| Indie | 327 | 8.2% |
| K-Pop | 278 | 7.0% |
| Afrobeats | 210 | 5.3% |
| Lo-fi | 201 | 5.0% |
| Classical | 166 | 4.2% |
| Country | 165 | 4.1% |
| Jazz | 132 | 3.3% |
Artists at the top
4,000 fictional listeners in view · Every country
See artist numbers
| Artist | Listeners | Share |
|---|---|---|
| SZA | 310 | 7.8% |
| Bad Bunny | 305 | 7.6% |
| The Weeknd | 302 | 7.5% |
| Dua Lipa | 298 | 7.4% |
| Taylor Swift | 294 | 7.3% |
| Travis Scott | 293 | 7.3% |
| Olivia Rodrigo | 290 | 7.2% |
| Tyler the Creator | 287 | 7.2% |
| Burna Boy | 285 | 7.1% |
| Drake | 285 | 7.1% |
The mood inside each genre
4,000 fictional listeners in view · Each square shows a genre’s mood mix
See mood and genre numbers
| Genre | Mood | Listeners | Within genre |
|---|---|---|---|
| Pop | Sad | 97 | 13.5% |
| Pop | Chill | 86 | 11.9% |
| Pop | Happy | 94 | 13.0% |
| Pop | Workout | 84 | 11.7% |
| Pop | Energetic | 98 | 13.6% |
| Pop | Sleep | 84 | 11.7% |
| Pop | Party | 83 | 11.5% |
| Pop | Focus | 95 | 13.2% |
| Hip-Hop | Sad | 76 | 12.6% |
| Hip-Hop | Chill | 73 | 12.1% |
| Hip-Hop | Happy | 84 | 13.9% |
| Hip-Hop | Workout | 77 | 12.7% |
| Hip-Hop | Energetic | 82 | 13.6% |
| Hip-Hop | Sleep | 62 | 10.2% |
| Hip-Hop | Party | 88 | 14.5% |
| Hip-Hop | Focus | 63 | 10.4% |
| Rock | Sad | 67 | 14.2% |
| Rock | Chill | 65 | 13.8% |
| Rock | Happy | 58 | 12.3% |
| Rock | Workout | 61 | 12.9% |
| Rock | Energetic | 53 | 11.2% |
| Rock | Sleep | 66 | 14.0% |
| Rock | Party | 46 | 9.7% |
| Rock | Focus | 56 | 11.9% |
| EDM | Sad | 46 | 12.0% |
| EDM | Chill | 53 | 13.9% |
| EDM | Happy | 56 | 14.7% |
| EDM | Workout | 54 | 14.1% |
| EDM | Energetic | 40 | 10.5% |
| EDM | Sleep | 49 | 12.8% |
| EDM | Party | 43 | 11.3% |
| EDM | Focus | 41 | 10.7% |
| R&B | Sad | 49 | 14.4% |
| R&B | Chill | 46 | 13.5% |
| R&B | Happy | 53 | 15.5% |
| R&B | Workout | 36 | 10.6% |
| R&B | Energetic | 39 | 11.4% |
| R&B | Sleep | 42 | 12.3% |
| R&B | Party | 34 | 10.0% |
| R&B | Focus | 42 | 12.3% |
| Indie | Sad | 45 | 13.8% |
| Indie | Chill | 45 | 13.8% |
| Indie | Happy | 42 | 12.8% |
| Indie | Workout | 44 | 13.5% |
| Indie | Energetic | 38 | 11.6% |
| Indie | Sleep | 31 | 9.5% |
| Indie | Party | 42 | 12.8% |
| Indie | Focus | 40 | 12.2% |
| K-Pop | Sad | 39 | 14.0% |
| K-Pop | Chill | 43 | 15.5% |
| K-Pop | Happy | 30 | 10.8% |
| K-Pop | Workout | 35 | 12.6% |
| K-Pop | Energetic | 27 | 9.7% |
| K-Pop | Sleep | 35 | 12.6% |
| K-Pop | Party | 34 | 12.2% |
| K-Pop | Focus | 35 | 12.6% |
| Afrobeats | Sad | 26 | 12.4% |
| Afrobeats | Chill | 31 | 14.8% |
| Afrobeats | Happy | 26 | 12.4% |
| Afrobeats | Workout | 28 | 13.3% |
| Afrobeats | Energetic | 26 | 12.4% |
| Afrobeats | Sleep | 22 | 10.5% |
| Afrobeats | Party | 33 | 15.7% |
| Afrobeats | Focus | 18 | 8.6% |
| Lo-fi | Sad | 23 | 11.4% |
| Lo-fi | Chill | 22 | 10.9% |
| Lo-fi | Happy | 26 | 12.9% |
| Lo-fi | Workout | 27 | 13.4% |
| Lo-fi | Energetic | 30 | 14.9% |
| Lo-fi | Sleep | 24 | 11.9% |
| Lo-fi | Party | 26 | 12.9% |
| Lo-fi | Focus | 23 | 11.4% |
| Classical | Sad | 26 | 15.7% |
| Classical | Chill | 15 | 9.0% |
| Classical | Happy | 19 | 11.4% |
| Classical | Workout | 25 | 15.1% |
| Classical | Energetic | 21 | 12.7% |
| Classical | Sleep | 29 | 17.5% |
| Classical | Party | 14 | 8.4% |
| Classical | Focus | 17 | 10.2% |
| Country | Sad | 22 | 13.3% |
| Country | Chill | 31 | 18.8% |
| Country | Happy | 12 | 7.3% |
| Country | Workout | 17 | 10.3% |
| Country | Energetic | 20 | 12.1% |
| Country | Sleep | 21 | 12.7% |
| Country | Party | 17 | 10.3% |
| Country | Focus | 25 | 15.2% |
| Jazz | Sad | 22 | 16.7% |
| Jazz | Chill | 19 | 14.4% |
| Jazz | Happy | 15 | 11.4% |
| Jazz | Workout | 11 | 8.3% |
| Jazz | Energetic | 18 | 13.6% |
| Jazz | Sleep | 12 | 9.1% |
| Jazz | Party | 16 | 12.1% |
| Jazz | Focus | 19 | 14.4% |
03 · Routine
How listening
fits the day.
Some profiles discover music through a weekly playlist, some download for offline listening, and some mix podcasts into the day. The chart shows how common each habit is in the current view; real research on why people listen to music adds mood, focus, identity, memory, and social connection.
Everyday listening habits
4,000 fictional listeners in view
See listening-habit numbers
| Habit | Listeners | Share |
|---|---|---|
| Discover Weekly | 2,280 | 57.0% |
| Offline mode | 1,710 | 42.8% |
| Podcasts too | 1,872 | 46.8% |
04 · Free versus paid
Paid does not mean
fewer skips.
Across the full set, Premium profiles skip slightly more often, not less. It is a useful reminder that a neat label does not guarantee a neat pattern—and it says nothing about real subscribers. Our Spotify Free versus Premium comparison keeps product features separate from personality claims.
Compare subscription plans
4,000 fictional listeners in view · Every plan
See subscription numbers
| Plan | Listeners | Average skip rate |
|---|---|---|
| Free | 1,820 | 27.7% |
| Premium | 1,193 | 28.6% |
| Family | 586 | 28.2% |
| Student | 401 | 28.7% |
05 · Models under pressure
A strong score can still tell
a weak story.
Two public notebooks tested the same 4,000 fictional profiles with prediction, statistical comparisons, and clustering. Their most useful result is not a new listener claim—it is a lesson in checking what a model score actually means.
What survives a closer check
Aggregate notebook findings · the same synthetic source used above
| The question | What we found | What we publish |
|---|---|---|
| Can a model predict songs per day? | The model scored 95.65%, while songs per day and listening minutes already correlate at 0.980. | No predictor. It mostly learns two versions of the same synthetic listening-intensity signal. |
| Does Premium reduce skipping? | Premium is 0.86 percentage points higher than Free; the 95% interval runs from -0.02 to 1.72, with d = 0.073. | Keep the comparison descriptive. The measured effect is negligible and says nothing about real subscribers. |
| Are Spotify and Apple Music listeners different? | All 4 comparisons have confidence intervals crossing zero; the largest absolute effect size is only 0.050. | No platform ranking or behavioral claim. |
| Did clustering discover listener types? | The three-cluster result repeats across random starts (mean ARI 0.993) but separates poorly (silhouette 0.081). | No generated personas and no connection to the listener quiz. |
Spotify vs Apple Music: the exact comparison
4 measures · Spotify minus Apple Music · fictional profiles only
| Measure | Spotify average | Apple Music average | Difference | Cohen’s d | 95% interval includes zero? |
|---|---|---|---|---|---|
| Daily listening minutes | 136.07 | 139.55 | -3.48 | -0.041 | Yes |
| Songs per day | 38.24 | 39.32 | -1.08 | -0.044 | Yes |
| Playlists | 9.01 | 9.04 | -0.03 | -0.011 | Yes |
| Skip rate (%) | 28.54 | 27.96 | 0.58 | 0.050 | Yes |
Every interval includes zero and every absolute effect size is below 0.05. Among these 4,000 fictional profiles, the measured Spotify and Apple Music differences are negligible.
Review the source notebooks: Music Streaming Habits 95% Beats and Music Streaming Habits 2026 .
Data literacy guide
How to read a music streaming
data analysis.
A model metric answers one narrow question. Reading correlation, uncertainty, effect size, cluster separation, and repeatability together prevents a high score from becoming a larger claim than the data supports.
What each statistic tells you
Plain-language interpretation · values from the notebook review
| Statistic | Question it answers | Result here | What it does not prove |
|---|---|---|---|
| Correlation | How closely do two numeric fields move together? | 0.980 between daily minutes and songs per day. | That one variable causes the other, or that the relationship exists in real listeners. |
| R² | How much target variation does the regression account for in its test data? | 95.65%, largely from overlapping listening-intensity fields. | That the model discovered a useful or generalizable behavioral rule. |
| 95% bootstrap interval | Which differences remain plausible after repeated resampling? | All four Spotify/Apple intervals include zero. | That the platforms are identical on every unmeasured behavior. |
| Cohen’s d | How large is a difference relative to the groups’ variation? | 0.073 for Premium versus Free; at most 0.050 for Spotify versus Apple Music. | Practical importance without context, representative sampling, and uncertainty. |
| Silhouette coefficient | How compact and separated are the clusters? | 0.081, close to the zero boundary between neighboring clusters. | That the groups are meaningful listener personas. |
| Adjusted Rand index | How similar are assignments across repeated cluster runs? | 0.993 mean ARI, so the assignments are repeatable. | That repeatable clusters are well separated, useful, or real. |
The key distinction is repeatability versus validity: an algorithm can consistently reproduce weak boundaries. That is why the cluster assignments stay out of the public listener quiz.
06 · A second dataset, tested
Can clean data still mislead you?
Absolutely.
We checked a separate 5,000-profile Kaggle file before turning it into another chart. It looks tidy, but the story underneath does not hold up. This audit is separate from the interactive sandbox above.
What the warning signs mean
5,000 profiles · no blanks · no repeated profiles once the ID is removed
| The claim | What we found | What it means |
|---|---|---|
| A trend from 2018–2024 | There is no year or date anywhere in the file. | There is no timeline to chart. |
| A fair platform comparison | 83.46% of profiles sit outside Spotify but still receive a Spotify-only Discover Weekly measure. | The discovery comparison is not credible. |
| A recommendation tool | The strongest relationship between any two numbers is just 0.038, with no documented sampling method. | There is no solid pattern to build on. |
| A practice dataset | The apparently generated file is complete, consistent, and released under CC0. | Useful for demos, not decisions. |
Clean columns do not automatically create trustworthy evidence. This file works for chart practice, schema exercises, and data-literacy demos—but not for market share, historical change, cause-and-effect claims, or listener recommendations. View the audited Kaggle source
Behind the story
What’s real here—
and what isn’t.
The interactive charts use version 1 of Music Streaming Habits 2026, updated 13 June 2026 and released under CC0: Public Domain. All 4,000 listener profiles are fictional.
The two notebook analyses use that same dataset, not additional listener evidence. We independently checked the public headline values and retained only aggregate findings—no notebook code, listener IDs, predictions, cluster assignments, or row-level derived output.
The final audit checks a different source: version 1 of Global Music Streaming Trends & Listener Insights, updated 9 March 2025 and released under CC0: Public Domain. None of its individual user IDs appear on this site; only the combined findings do.
Most percentages use the profiles currently in view. The plan comparison keeps all subscriptions visible, while the mood chart shows how each genre’s listeners divide across moods.
Neither file represents a real country, platform, age group, or subscription audience. Use the page to explore how charts react—not to make claims about people.
Quick answers
Music streaming habits
dataset FAQ.
These answers separate what the synthetic dataset demonstrates from what would require representative, documented listener research.
Is the Music Streaming Habits 2026 dataset real?
No. It contains 4,000 fictional listener profiles created for learning and visualization. Its countries, platforms, subscriptions, genres, moods, and behaviors do not come from a survey or records from real streaming services.
What does the 95.65% music prediction score mean?
The gradient-boosting regression predicts synthetic songs per day with an R² of 0.9565. Most of that performance is explained by daily listening minutes, a field already correlated at 0.980 with songs per day, so the score is not evidence of a new behavioral discovery.
Do Premium listeners skip fewer songs than Free listeners?
Not in this file. Premium profiles average 28.61% and Free profiles 27.75%, a difference of 0.86 percentage points. The effect size is negligible (d = 0.073), and the 95% interval includes zero.
Can K-means clusters become music-listener personas?
Not responsibly from this analysis. The assignments repeat across runs, but the silhouette score of 0.081 shows poor separation. Stable labels do not turn overlapping synthetic groups into real audience segments.
Continue exploring
From a data demo
to real listening research.
These five articles keep real surveys, platform features, recommendation research, and listening motives separate from the fictional profiles above.
Make it personal
What kind of
listener are you?
Answer six quick questions for a playful listening profile and a place to start in the Doldur catalog.
Take the listener quiz