Claim Missing Document
Check
Articles

Found 1 Documents
Search

Penerapan Algoritma K-Means untuk Klasterisasi Lagu Terpopuler 2025 Versi Spotify Al Aini, Nuril Aliya; Endynda, Rahmadanix Cinta; Al Rosyid, Harun
Jurnal Sains Dan Teknologi | E-ISSN : 3063-9980 Vol. 2 No. 3 (2026): Januari - Maret
Publisher : GLOBAL SCIENTS PUBLISHER

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

This study aims to analyze and group the most popular songs on Spotify in 2025 based on the emotions or moods they represent. The research uses the K-Means algorithm. The data includes the top 3,000 songs with ten main audio features: valence, energy, danceability, tempo, loudness, acousticness, instrumentalness, speechiness, liveness, and duration_ms. The research process includes several stages, such as data preprocessing, normalization using the standardization method, determining the best number of clusters with the Elbow and Silhouette Score methods, and applying the K-Means algorithm. The results show that the best number of clusters is four. The first cluster includes 1,180 songs that represent happy or cheerful moods, the second cluster consists of 930 songs that are energetic, the third cluster has 260 songs that show calm or relaxed feelings, and the fourth cluster includes 630 songs with sad or mellow tones. The results are visualized using PCA and heatmap, showing clear and accurate differences between the clusters based on audio features. These findings indicate that the K-Means algorithm is effective for analyzing the mood of songs and can be used in developing music recommendation systems based on mood.