For my final project, I chose to visualize different characteristics of music found on Spotify to explore the relationships between them using the Spotify Tracks Dataset from Kaggle, found here.
You can find information on every field in the dataset at the provided link; here is a description of each field used in the project:
This is a heatmap that compares the popularity of every key within different genres. For example, in the top left, every acoustic song in the key C, regardless of mode, have a combined popularity score of 6,053. Within the dataset, the two most common keys are C and G, which appear to be the two most popular keys, as well. It's understandable that these keys are used most often in music, as they are two of the simplest to play.
This scatterplot explores correlations between valence and danceability in a song. Naturally, one would expect a danceable song to be considered positive, or having a high valence score. This trend is generally observed within the plot, as the points trend upwards and to the right as you continue through the graph with a few outliers. Tempo range is a calculated field I made to make it easier to categorize datapoints by tempo.
This treemap compares every genre by instrumentalness first, then energy. The size of each square is the sum of instrumentalness within the genre, and the color is the sum of energy. I found it interesting to see that there is no obvious correlation between instrumentalness and energy across genres.
This is a bar chart measuring the sum of liveness of each genre. I decided to do this visualization to see if there were any genres on Spotify that were recorded live more than others, and as you can see, there are a few! Comedy, pagode, samba, and sertanejo appear to have the most live recordings uploaded to Spotify.
This is a story of three different bar charts, all measuring a different value by key and mode. When first looking at this dataset, I had a feeling that there would be a correlation between valence, popularity, and danceability. I explored the relationship between valence and danceability in the scatterplot above, but I wanted to see how they would compare to one another from a bar chart perspective and when considered alongside mode and key. As you can see when tabbing through the story, there is little to no change between the charts. I expected this, as one could infer that a song that is considered more positive, or to have a higher valence, would be more danceable; if a song is more danceable, it's likely that more people will listen to the song, boosting its popularity. I found it very interesting that in the key of B, minor songs are considered more popular, danceable, and have a higher valence, which is the oppostite of what I would expect when comparing these across modes.
