File size: 3,235 Bytes
e95ea95
 
 
3ab5c68
e95ea95
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3ab5c68
e95ea95
 
 
 
 
 
3ab5c68
e95ea95
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3ab5c68
e95ea95
 
 
 
 
 
 
 
 
 
 
 
 
3ab5c68
 
 
e95ea95
 
 
3ab5c68
e95ea95
 
 
 
 
 
 
 
 
 
 
 
 
3ab5c68
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
# Assignment #1 - EDA & Dataset - orian rivlin 

**Goal:**
Explore which audio features are most strongly related to a track’s popularity on Spotify.
This repository includes the dataset sample, a well-documented notebook, saved figures, and a short video walkthrough.

---

## Dataset
- **Name:** Spotify Features (Sample)
- **File:** `SpotifyFeatures_sample.csv`
- **Rows:** ~10,000  |  **Columns:** 18 (mostly numeric)
- **Target:** `popularity` (0–100)
- **Main numeric features:** `danceability`, `energy`, `loudness`, `speechiness`, `acousticness`, `instrumentalness`, `liveness`, `valence`, `tempo`, `duration_ms`
- **Source:** Spotify Tracks DB (via Kaggle)

**Main question:** Which audio features are most strongly related to popularity?

---

## Data Cleaning (Step 2.1)
- Removed duplicate rows.  
- Checked missing values (none in key columns)
- Dropped rows with NA in core numeric features if present.  
- Built a numeric-only DataFrame for stats and correlations.

Result: 10,000 rows × 11 numeric features** ready for EDA.

---

## EDA Highlights

### 1) Distributions
- **Popularity** is concentrated around 40-60 with few very high hits (80+).  
- **Danceability** ~ bell-shaped around ~0.56.  
- **Energy** skews higher (many tracks at 0.7-0.9).  
- **Valence** is broadly spread (neutral on average).

![Popularity distribution](materials/popularity_hist.png)
![Feature distributions](materials/features_dists.png)

### 2) Correlations with Popularity
Top relationships:

| Feature          | Corr. with popularity |
|------------------|-----------------------|
| **loudness**     | **+0.31**             |
| **energy**       | **+0.27**             |
| danceability     | +0.06                 |
| tempo            | +0.02                 |
| valence          | −0.06                 |
| **acousticness** | **−0.35**             |

**Interpretation.** Popular tracks tend to be louder and more energetic, and less acoustic.

![Correlation heatmap](materials/corr_heatmap.png)

### 3) Popular vs. Unpopular (Top 10% vs Bottom 10%)
- **Energy** and (slightly) **danceability** are higher among top-10% tracks.  
- **Tempo** shows little difference.  
- **Valence** is only slightly higher for top tracks.

![Top vs Bottom 10%](materials/top_vs_bottom_boxplot.png)

---

## Key Insights
- Popularity on Spotify is uneven: only a small minority of tracks become very popular.  
- Loudness and energy are the strongest positive correlates of popularity; acousticness is the strongest negative correlate.  
- Highly popular songs tend to sound modern/produced (loud, energetic), while purely acoustic/instrumental tracks underperform on average.

---

## Prestation Video 
**Video:** (https://youtu.be/CIxhSgZXnyw)

---

## Files
- `SpotifyFeatures_sample.csv` – dataset sample  
- `spotify_eda_notebook.ipynb` – code & plots  
- `materials/` – exported materials used in this README  
- `README.md` – this summary

---

## Notes and Decisions
- **Outliers** (very long tracks; very low/high popularity) were kept as real observations; statistics were interpreted cautiously.  
- Correlations are modest overall (music success is multifactorial); results describe associations, not causation.