Jenny Ho




Making property data approachable at Archipelago


Improving access to social services at Healthify


Personal projects

Personal projects



What do the 100 greatest metal albums sound like?
Data analysis and visualization



This data visualization project is about Rolling Stone’s 100 Greatest Metal Albums and how they “sound” based on Spotify’s API.

I’m a casual fan of metal music, so this dataset was a fun way to apply what I learned from data visualization classes at Parsons. Data cleansing, analysis, and visualization done with Tableau. Visual design in Figma.

View the project here.


What does musical data look like?

The dataset contains two sheets with different levels of granularity.
  • The first has subjective, qualitative data about each album. Examples: rankings, ratings, subgenres.
  • The second is about songs on the albums. It breaks them down into measurable audio attributes from Spotify. Examples: beats per minute, track order, energy level, and danceability scores.

I started with an exploratory data analysis in Tableau for albums and songs to find a few potential story ideas.

Album trends and patterns.
Song trends and patterns.

What makes an album great?

 
I arrived at at few questions that I wanted to answer. 
  • What does the overall structure of an album look like, from the first track to the last?
  • What do the top ranked and the most popular albums look like?
  • Which factors affect Rolling Stone rankings and Spotify popularity? 

Starting with sketches.
Ending with options for stories and wireframes.

The dataset needs fixing up. 

A few things to clean up before designing:
  • Combine the two sheets into one so that I can aggregate audio attributes by album and track number.
  • Correctly calculate song duration because it was misclassified as date/time.
  • Find any outliers and check for errors. A live album was accidentally included in the dataset — it had too many tracks compared to a typical studio album.

Cleaning data in Tableau Prep.
Combined album and song data.

It has to look metal enough.

On top of a simple design system, I added details to make it obvious this project is about metal music. This includes fun (but readable) typefaces, dark colors, and album art as part of the visualization.

View the project here.

Introducing the dataset.
Aggregating the 100 albums.
Pick out a few examples.

Factors influencing rank and popularity.