adiprog14's picture
Update README.md
5c064fc verified
|
Raw
History Blame Contribute Delete
2.39 kB
# Spotify Tracks Dataset - EDA Analysis
**Student:** Adi Toledano
**Dataset Source:** maharshipandya/spotify-tracks-dataset
**Goal:** Analyze what features influence a song's popularity on Spotify
<video src="https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/presentation%20%231.mp4" controls="controls" style="max-width: 720px;"></video>
---
## Dataset Overview
- **Rows:** 113,397 songs
- **Columns:** 19 features
- **Source:** HuggingFace - maharshipandya/spotify-tracks-dataset
---
## Research Questions & Key Findings
### 1. What is the distribution of song popularity?
- Average popularity: 33.3
- Most songs have low popularity, few reach the top
![Popularity Distribution](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/popularity_distribution.png)
---
### 2. Which genres are most popular?
- 🥇 Pop-Film: 59.3
- 🥈 K-Pop: 57.0
- 🥉 Chill: 53.7
![Top Genres](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/top_genres.png)
---
### 3. Do danceability and energy affect popularity?
- Danceability correlation: 0.034 (almost none)
- Energy correlation: -0.000 (none)
- Surprising finding: audio features barely affect popularity!
![Danceability and Energy](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/danceability_energy.png)
---
### 4. Strong correlations found:
- Energy ↔ Loudness: 0.76
- Energy ↔ Acousticness: -0.73
- Danceability ↔ Valence: 0.48
![Correlation Heatmap](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/correlation_heatmap.png)
---
### 5. Are explicit songs more popular?
- Explicit songs: 36.5 average
- Non-explicit songs: 33.0 average
- Explicit songs are slightly more popular
![Explicit Popularity](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/explicit_popularity.png)
---
### 6. Who are the most popular artists?
- 🥇 Sam Smith & Kim Petras: 100
- 🥈 Bizarrap & Quevedo: 99
- 🥉 Manuel Turizo: 98
![Top Artists](https://huggingface.co/datasets/adiprog14/spotify-eda-analysis/resolve/main/top_artists.png)
---
## Conclusion
Genre and artist identity matter more than audio features
when it comes to a song's popularity on Spotify!
---
## Files
- `spotify_eda.ipynb` - Full analysis notebook
- `spotify_cleaned.csv` - Cleaned dataset