I often use Spotify to wake my kids up in the morning and try to choose songs that I would consider to be high valence and high energy. I plan to investigate if Spotify track popularity is linearly correlated to valence and energy and made the visualization below to get a sense of how these variables interacted.

I looked at the distribution of each variable to create 3 levels each for valence, energy and popularity. I was surprised to see just how few songs out of the sample have high values in all three areas. This visualization got me even more interested in the few songs that have high scores across all three variables and in my future analysis. I am also curious to investigate the few songs that fall into the high valence, high popularity and low energy category. What is clear is the sheer number of tracks that are classified as high/medium valence and high/medium energy that do not have a high value for track popularity. The category with the highest number of tracks fall into the Medium Valence, High Energy and Low Popularity bin and I hypothesize now to find a negative linear relationship between valence and energy and track popularity.