orianrivlin commited on
Commit
3ab5c68
·
verified ·
1 Parent(s): c0cbddd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -10
README.md CHANGED
@@ -1,7 +1,7 @@
1
  # Assignment #1 - EDA & Dataset - orian rivlin
2
 
3
  **Goal:**
4
- Explore which audio features are most strongly related to a track’s **popularity** on Spotify.
5
  This repository includes the dataset sample, a well-documented notebook, saved figures, and a short video walkthrough.
6
 
7
  ---
@@ -24,14 +24,14 @@ This repository includes the dataset sample, a well-documented notebook, saved f
24
  - Dropped rows with NA in core numeric features if present.
25
  - Built a numeric-only DataFrame for stats and correlations.
26
 
27
- Result: **10,000 rows × 11 numeric features** ready for EDA.
28
 
29
  ---
30
 
31
  ## EDA Highlights
32
 
33
  ### 1) Distributions
34
- - **Popularity** is concentrated around **40-60** with few very high hits (80+).
35
  - **Danceability** ~ bell-shaped around ~0.56.
36
  - **Energy** skews higher (many tracks at 0.7-0.9).
37
  - **Valence** is broadly spread (neutral on average).
@@ -51,7 +51,7 @@ Top relationships:
51
  | valence | −0.06 |
52
  | **acousticness** | **−0.35** |
53
 
54
- **Interpretation.** Popular tracks tend to be **louder** and **more energetic**, and **less acoustic**.
55
 
56
  ![Correlation heatmap](materials/corr_heatmap.png)
57
 
@@ -65,13 +65,13 @@ Top relationships:
65
  ---
66
 
67
  ## Key Insights
68
- - Popularity on Spotify is **uneven**: only a small minority of tracks become very popular.
69
- - **Loudness** and **energy** are the strongest positive correlates of popularity; **acousticness** is the strongest negative correlate.
70
- - Highly popular songs tend to sound **modern/produced** (loud, energetic), while purely acoustic/instrumental tracks underperform on average.
71
 
72
  ---
73
 
74
- ## 2-3 Minute Video
75
  **Video:** (https://youtu.be/CIxhSgZXnyw)
76
 
77
  ---
@@ -85,5 +85,5 @@ Top relationships:
85
  ---
86
 
87
  ## Notes and Decisions
88
- - **Outliers** (very long tracks; very low/high popularity) were **kept** as real observations; statistics were interpreted cautiously.
89
- - Correlations are **modest** overall (music success is multifactorial); results describe associations, not causation.
 
1
  # Assignment #1 - EDA & Dataset - orian rivlin
2
 
3
  **Goal:**
4
+ Explore which audio features are most strongly related to a track’s popularity on Spotify.
5
  This repository includes the dataset sample, a well-documented notebook, saved figures, and a short video walkthrough.
6
 
7
  ---
 
24
  - Dropped rows with NA in core numeric features if present.
25
  - Built a numeric-only DataFrame for stats and correlations.
26
 
27
+ Result: 10,000 rows × 11 numeric features** ready for EDA.
28
 
29
  ---
30
 
31
  ## EDA Highlights
32
 
33
  ### 1) Distributions
34
+ - **Popularity** is concentrated around 40-60 with few very high hits (80+).
35
  - **Danceability** ~ bell-shaped around ~0.56.
36
  - **Energy** skews higher (many tracks at 0.7-0.9).
37
  - **Valence** is broadly spread (neutral on average).
 
51
  | valence | −0.06 |
52
  | **acousticness** | **−0.35** |
53
 
54
+ **Interpretation.** Popular tracks tend to be louder and more energetic, and less acoustic.
55
 
56
  ![Correlation heatmap](materials/corr_heatmap.png)
57
 
 
65
  ---
66
 
67
  ## Key Insights
68
+ - Popularity on Spotify is uneven: only a small minority of tracks become very popular.
69
+ - Loudness and energy are the strongest positive correlates of popularity; acousticness is the strongest negative correlate.
70
+ - Highly popular songs tend to sound modern/produced (loud, energetic), while purely acoustic/instrumental tracks underperform on average.
71
 
72
  ---
73
 
74
+ ## Prestation Video
75
  **Video:** (https://youtu.be/CIxhSgZXnyw)
76
 
77
  ---
 
85
  ---
86
 
87
  ## Notes and Decisions
88
+ - **Outliers** (very long tracks; very low/high popularity) were kept as real observations; statistics were interpreted cautiously.
89
+ - Correlations are modest overall (music success is multifactorial); results describe associations, not causation.