The goal of this project is to explore the characteristics of songs that became hits on Spotify in 2023. I asked: Which artists dominated? How do audio features like energy, danceability, or valence relate to popularity? How does song performance vary across different platforms?
You can check out part 2 here where I use Python to dig deeper!
Understanding what makes a hit is valuable not only for music fans, but also for artists, record labels, and marketers who want to know which qualities help songs stand out across streaming platforms.
By combining SQL, Python, and Tableau, I was able to extract, clean, analyze, and visualize real-world music data to answer these questions.
Dataset Overview
The dataset comes from Kaggle’s Spotify Top 2023 dataset. It contains features such as:
- Song and artist information
- Stream counts
- Chart presence across Spotify, Apple Music, Deezer, and Shazam
- Audio features (energy, danceability, valence, etc.)
This combination makes it a great dataset to study both artist dominance and the attributes of songs that cross platforms.
Tools Used for Step 1
- Database: MySQL
- Skills demonstrated: Table creation, data aggregation, filtering, grouping, ordering, ranking, column renaming
The entire dataset:
Before we begin, it’s important to take a look at the entire dataset as we consider the problem/problems we are solving.
SELECT *
FROM spotify2023;
1. SQL: Data Extraction & Aggregation
- Purpose: Summarize raw song data to answer high-level questions.
- Key Tasks & Questions:
- Identify artists with the most songs in the Spotify top charts.
- Calculate average streams per artist and per platform.
- Count cross-platform hits across Spotify, Apple, Deezer, and Shazam.
- Skills Demonstrated: Table creation, data aggregation, filtering, grouping, ordering, and ranking.
Songs that had the most streams:
— Which songs had the most streams?
SELECT track_name, streams
FROM spotify2023
ORDER BY streams DESC;
The song Circles had the most streams followed by Levitating (feat. DaBaby). These results highlight which individual tracks captured the most global attention on Spotify in 2023.
Column cleanup for readability
Renaming improves clarity when sharing queries with collaborators or stakeholders.
— Renaming to make it easier for reading the data
ALTER TABLE spotify2023
RENAME COLUMN artist(s)_name TO artist_name;
— Which artists have multiple songs in the top list?
SELECT artist_name, COUNT(*) AS track_count
FROM spotify2023
WHERE in_spotify_charts > 0 OR in_apple_charts > 0 OR in_deezer_charts > 0 OR in_shazam_charts > 0
GROUP BY artist_name
ORDER BY track_count DESC;
As you can see Taylor has the most tracks in the top 10! Bad bunny comes in second, and SZA in third. This shows how consistently these artists released songs that connected with listeners across platforms.
— Which songs charted across multiple platforms (Spotify, Apple, Deezer, Shazam)?
SELECT track_name, IF(in_spotify_charts > 0, 1, 0) + IF(in_apple_charts > 0, 1, 0) + IF(in_deezer_charts > 0, 1, 0) + IF(in_shazam_charts > 0, 1, 0) AS in_charts
FROM spotify2023
WHERE IF(in_spotify_charts > 0, 1, 0) + IF(in_apple_charts > 0, 1, 0) + IF(in_deezer_charts > 0, 1, 0) + IF(in_shazam_charts > 0, 1, 0) > 1;
This lets us see which songs charted across multiple platforms including Spotify, Apple, Deezer, and Shazam. Seven, LALA, vampire, Cruel Summer, and others are in all four platforms. Cross-platform hits indicate the songs with the broadest listener appeal, not just on Spotify but across different user bases.
— Artists with highest total streams
SELECT
artist_name, SUM(streams) as sum_streams
FROM spotify2023
GROUP BY artist_name
ORDER BY sum_streams DESC
LIMIT 10;
Who do you think is the artist with the highest total streams? If you guessed Taylor Swift, you’re correct. Bad Bunny and The Weeknd come in second and third respectively.
Thanks for checking out Part 1 of my project! In this step, I used SQL to explore artist dominance, cross-platform hits, and streaming leaders.
👉 Next up in Part 2 (Python), I’ll dig into audio features like danceability, valence, and energy to uncover patterns in what makes a song popular. After that, in Part 3 (Tableau), I’ll create interactive dashboards to visualize these trends.
Lindsay