As I was brainstorming ideas and running through databases for cool json and CSV datasets I came across this really cool website called Organise Your Music. What it does is, it establishes a connection with your Spotify account and runs a script that collects information about what kind of genre music is in your libray as well as a multidue of different characteristics, such as energy, valence, danceability etc. about each song.
As you can see, pop music seems to dominate my library with rock following relatively close by. This is kinda funny because I thought that I didn't listen to too much pop music in my free time. However, this is not indicative of how much time I have actually listened to one genre or the other as I think disco and funk, although presented in much lesser numbers take up much more of my listening time than funk. This graph also doesn't look at the country of origin of the song or the language in which it is sung in. I can promise you, that it isn't predominantly English.
First some terminology:
- Energy - The energy of a song - the higher the value, the more energtic the song
- Danceability - The higher the value, the easier it is to dance to this song
- Valence - The higher the value, the more positive mood for the song.
NB! All of these variables are measured on scale of 0-100
Looking at the different data I could gather from Organize Your Music I came across the aforementioned indicators. I had a hunch that there could be a connection between the energy and the danceability and valence of the song. Thus, I plotted a scatterplot of energy against danceability (Blue) and valence (red). To my demise the distribution came out as a relativly random distribution for both variables. I further calculated the r-squared value for both variables which for danceability came to 0,0321 and for valence came at 0,048. This indicates that the model can explaint to a small extent the positive correlation between enerygy against valence and danceability, but nothing significant.