GitHub repository for the Data and Results Visualization course project @ Politecnico di Milano. The goal of the course was to provide the principles and rules of Data Analysis and Visualization.
Our project studies the impact of COVID-19 on the music industry with particular focus on music streamings. We gathered our data using the Spotify API and the Spotify top 200 weekly. We also used the iTunes API to retrieve genre information for each song.
The PDF export of the project can be found here: THE_IMPACT_OF_COVID-19_IN_MUSIC_INDUSTRY.pdf
Our repository contains 3 different notebooks + data folders.
Data folder are organized in nation and nation_features, for example 'it' which stands for Italy and it_features. The first type of folders contain the plain top 200 weekly chart of the mentioned nation, spanning from the first available data (which is different for every nation) until August 2020.
The second type of folders contain the top 200 weekly with features, such as Danceability, Energy ecc... These features have been retrieved using Spotify API. Music genres have been collected using iTunes API.
NB: music genre labels need to be checked before being used, some songs are not present inside the iTunes database and/or have different song titles, thus some music genre labels might be missing or wrong.
For our project, we manually checked if the labels were correct (limited to the data we used), thus we can't guarantee that all the genre labels are available and correct.
Data Download allows the user to download all the top 200 charts in the selected time span for the selected nations. Once the files containing the "plain" top 200 have been downloaded, Download features is used to extract and store features for each song.
THE_IMPACT_OF_COVID-19_IN_MUSIC_INDUSTRY.ipynb contains the jupyter notebook of our project. We exported the notebook to PDF format in THE_IMPACT_OF_COVID-19_IN_MUSIC_INDUSTRY.pdf, so you will just need a PDF reader to open it. If you're not a Python user/developer and you're just interested in our conclusion and data analysis, just go on with the PDF file. Notice that (obviously) interactive visualization are not rendered and there are some formatting problems due to the conversion from jupyter notebook to PDF. Beside this, the plots and the analysis still hold.
We don't own any rights on the data, it has been downloaded using public available APIs and dataset from Spotify and iTunes. If you would like to contribute to improve the data quality or increase its quantity, feel free to do it and contact me.