Description
This project explores music popularity dynamics through Shazam-related signals over time. The objective was to move from static rankings to temporal understanding: trend acceleration, periodic effects, anomalies, and momentum phases.
Key Features
- Temporal Exploration: Built a full time-series pipeline to inspect trend evolution instead of one-shot snapshots.
- Seasonality & Patterns: Identified recurring patterns and shifts in listening behavior.
- Forecasting Logic: Evaluated predictive baselines to estimate short-term movement.
- Insight Visualization: Produced clear visual narratives to communicate trend changes and signal quality.
Technologies Used
- Python: End-to-end analysis pipeline from ingestion to visualization.
- Pandas: Time-series data manipulation and trend exploration.
- NumPy: Statistical computations and signal processing.
- Plotly: Interactive trend visualizations and temporal diagnostics.
- Jupyter: Iterative experimentation and insight documentation.
My Role
I did this project solo, end to end. From scraping and cleaning the Shazam data to building the time-series features and designing the visual narrative, every step was mine. It was also an excuse to get better at explaining temporal patterns without losing the reader in jargon.
What I Learned
Time-series work is humbling. The data always has a story it wants to tell, and your job is to listen before you force a model on it. I also learned that a good trend chart is worth more than a dozen accuracy metrics when presenting to non-technical audiences.
Outcome
The project provided a robust framework to interpret Shazam dynamics over time, helping transform raw events into actionable insights on music trajectory and trend behavior.