This project investigates the dynamics of the global music market, analyzing trends and factors influencing popular songs. Using two Spotify datasets, one with historical songs (up to March 2020) and another with popular recent tracks, our team explored how music characteristics have evolved and identified new methods to categorize songs based on their attributes.
- Popularity Trends: From 2010 to 2020, the percentage of popular songs consistently ranged between 15% and 20% of the total, averaging at 17%.
- Duration & Genre Evolution: Song durations have decreased, while genre diversity within individual tracks has grown, challenging traditional genre-based classification methods.
- Pre- vs. Post-COVID Analysis: While danceability and acousticness have slightly increased, the overall characteristics of popular songs remain stable across these periods.
- Revised Categorization: By clustering songs using "valence" and "energy," we identified four distinct groups:
- Low Energy, Low Valence (Most Popular)
- Low Energy, High Valence
- High Energy, High Valence
- High Energy, Low Valence (Least Popular)
- Data Cleaning: Before analysis, we addressed missing or inconsistent data. This included removing duplicate entries, handling missing values using imputation where appropriate, and standardizing the formatting of song attributes.
- Exploratory Data Analysis (EDA): Key metrics and trends were visualized to understand historical patterns and recent shifts.
- Clustering Analysis: K-means clustering using the variables "valence" and "energy" revealed hidden subsets of songs, offering a novel categorization approach.
- Statistical Modeling: We incorporated the new clusters into predictive models, optimizing accuracy and deriving actionable insights for stakeholders.
- Music Producers: Insights on clustering can guide creative decisions, targeting popular characteristics.
- Industry Collaborations: Organizations can better understand audience preferences and refine marketing strategies.
- DSO 528 Final Presentation_Team6.pdf: Comprehensive project presentation.
- DSO 528 Final Project Executive Summary.pdf: Condensed summary of the project’s objectives, findings, and conclusions.
- Review the "Executive Summary" for a quick overview.
- Reference the "Presentation" for visual summaries and graphs.
This project was completed as part of the coursework for DSO 528, leveraging collaborative efforts and Spotify’s datasets. We extend our gratitude to the course instructors and team members for their contributions.
This project is released under the MIT License.