Spotify-Recommendation-Algorithms-

This project builds a simple and effective music recommendation system using Spotify song data and the K-Nearest Neighbors (KNN) algorithm.

1. Import Libraries

• Essential libraries are imported:
• pandas, numpy for data manipulation.
• sklearn for machine learning tools (e.g., MinMaxScaler, NearestNeighbors).
• warnings to suppress unnecessary warnings.

We'll use content-based filtering using audio features (like tempo, energy, danceability, etc.) to recommend similar songs.

Singer	Acousticness	Danceability	Energy	Track ID	Instrumentalness	Key	Liveness	Loudness	Mode	Song Name	Speechiness	Tempo	Spotify Link	Valence
Aastha Gill	0.485	0.770	0.824	39ujbBjTwwqUFySaCYDMMT	0.000000	1	0.318	-6.491	0	Proper Patola	0.0851	172.006	Link	0.490
Aastha Gill	0.143	0.825	0.666	5cjVsWqIkBQC7acTRhL0RO	0.000003	4	0.237	-4.847	0	Kamariya	0.0554	96.987	Link	0.763
Aastha Gill	0.236	0.663	0.551	3XYvdqcZrTmRntFDDbJkJd	0.000036	3	0.0923	-8.272	0	Buzz (feat. Badshah)	0.1090	113.314	Link	0.601
Aastha Gill	0.00323	0.919	0.571	46GBoFCdFZZSjuGaZjZmGv	0.001680	5	0.103	-7.175	0	Saara India	0.0687	105.007	Link	0.231
Aastha Gill	0.129	0.867	0.720	6VwVEIiCro1EMyh9B6Om3v	0.000000	9	0.228	-5.188	0	Drunk n High	0.0619	104.974	Link	0.755

2. Load the Dataset

• Dataset: SingerAndSongs.csv
• Contains ~songs with metadata and audio features (e.g., Energy, Tempo, Danceability).
• data = pd.read_csv("SingerAndSongs.csv", encoding='latin1')

Explore the Dataset Structure Run this code to understand what columns you have:

Feature	Count	Mean	Std Dev	Min	25%	Median	75%	Max
Acousticness	2231	0.3962	0.2963	0.0003	0.1180	0.3440	0.6475	0.9940
Danceability	2231	0.6127	0.1560	0.1560	0.5050	0.6220	0.7340	0.9710
Energy	2231	0.6582	0.1900	0.1030	0.5200	0.6690	0.8155	0.9880
Instrumentalness	2231	0.0120	0.0784	0.0000	0.0000	0.0000	0.0002	0.9670
Key	2231	5.3967	3.4597	0.0000	2.0000	6.0000	8.0000	11.0000
Liveness	2231	0.1841	0.1418	0.0222	0.0948	0.1260	0.2455	0.9720
Loudness (dB)	2231	-7.2509	2.9299	-20.0900	-8.8415	-6.8870	-5.3325	0.0030
Mode	2231	0.5737	0.4946	0.0000	0.0000	1.0000	1.0000	1.0000
Speechiness	2231	0.0749	0.0696	0.0232	0.0342	0.0483	0.0814	0.6840
Tempo (BPM)	2231	115.489	26.2875	55.8320	95.9870	109.9820	130.0980	214.0160
Valence	2231	0.5584	0.2163	0.0394	0.3850	0.5630	0.7310	0.9700

3. Data Cleaning

• Drops irrelevant columns: 'Index', 'Title', 'Artist', 'Top Genre', 'Year'
• Keeps only numeric/audio feature columns.

Available columns: ['Singer', 'acousticness', 'danceability', 'energy', 'id', 'instrumentalness', 'key', 'liveness', 'loudness', 'mode', 'Song name', 'speechiness', 'tempo', 'track_href', 'valence'] Required columns 'track_name' and 'artist_name' not found. First few column names: Index(['singer', 'acousticness', 'danceability', 'energy', 'id'], dtype='object')

Danceability	Energy	Loudness	Speechiness	Acousticness	Instrumentalness	Liveness	Valence	Tempo
0.753	0.815	0.677	0.0937	0.488	0.000000	0.311	0.484	0.734
0.821	0.636	0.759	0.0487	0.144	0.000003	0.226	0.778	0.260
0.622	0.506	0.588	0.1298	0.237	0.000037	0.074	0.603	0.363
0.936	0.529	0.643	0.0689	0.003	0.001737	0.085	0.206	0.311
0.872	0.697	0.742	0.0586	0.130	0.000000	0.217	0.769	0.311

Build Recommendation Model

• Uses NearestNeighbors from sklearn to find songs with similar audio features.
• Fits the model on scaled feature data.

Define the Recommendation Function & Test the Recommender

• ✅ Found: Paani Paani
• ✅ Found: Naagin
• ✅ Found: 52 Non Stop Dilbar Dilbar Remix(Remix By Kedrock,Sd Style)
• ✅ Found: Dj Waley Babu (feat. Aastha Gill)
• ✅ Found: Crazy Lady
• ✅ Found: Badtameez
• ✅ Found: Call Waiting - Reprise
• ✅ Found: Light Kardo Band (feat. Aastha Gill)
• ✅ Found: Kareja Kareja (feat. Aastha Gill)

Visualize Recommendations

Average similarity (accuracy-like score): 99.65%

The plot_similar_songs() function then plots these recommendations in a line graph.

📝 Conclusion

• This dataset provides audio features for 2,231 tracks, highlighting various attributes like danceability, energy, acousticness, and valence. Key takeaways include:
• Tracks vary widely in energy and danceability, with a mix of acoustic and electronic styles.
• Valence indicates a generally positive mood, while tempo spans from slow to fast tracks.
• Missing columns (track_name and artist_name) can be mapped from available data.

---

📄 License

This project is protected under copyright © Md Altamash Alam, 2025.

All rights reserved. Unauthorized copying, distribution, modification, or use of any part of this project without explicit permission is strictly prohibited.

If you wish to use or reference any part of this project for academic, personal, or commercial purposes, please contact the author for permission.

---

© Md Altamash Alam, 2025 – All Rights Reserved.