Pol Serramalera
Data Analytics full time 09/nov/2020
This project seeks to create a program that returns a random song froma 10000 song cluster based on the audiofeatures of the song inputed by a user. The program will return the song and artist of a similar song in terms of acousticness, danceability, keys, tempo etc.
Created on the code:
·billboard.csv ·final_playlist.csv
- First I have done a little of webscrapping in order to get the billboard top 100 hot songs from the billboard website
- Then definining a function to convert any spotify playlist into a datafram with the audiofeatures of eachsong of the playlist
- With the function defined, extract audio fetures of 10000 songs
- Clustering the dataframe with kmeans
- Defining the loop function so the user can easily input a song and get a recommendation back