About Our Project

Here is a brief description of our project. If you'd like to check the project itself, follow the link.

We extracted our information using Spotify API which gave us access to the data on songs. We did that by first inputting the unique ID into the API. After creating a playlist with wanted Eurovision songs we exported it via Spotify’s automated feature which returned JSON files containing the information about the artist, song, name, year, popularity and various audio features (danceability, energy, loudness, speechiness, acousticness, instrumentalness, liveness, valence, tempo, duration (in milliseconds) and a time signature). The main audio features we focus on are danceability, energy, loudness, acousticness, valence, tempo. With this information we created a dataset containing information about participating artists and their songs, grouped by the year in which their song was representing their country. After that we could start analysing the songs’ audio features.

From the analysis we performed using visualisation tools (Flourish) we came to the conclusion that there are a few existing trends in the audio featurs of Eurovision songs. The latter are created with a high focus on danceability, energy and loudness. There is a general preference for the less acoustic songs (with the exception of a few songs) and a relative unbiasedness to the positive (happy, cheerful, euphoric) or negative (sad, depressed, angry) songs. The trends described above are best observed in the recent songs from 2020 and 2021. Overall, we noticed that in the past decade the performing songs have managed to keep relatively constant audio features.