Explore the behaviors of Music Artists through Twitter Data
By Shrishti Charan Wahal, Mehar Jossan, Lakshaya Gupta and Mandeep Kaur · May 5,2016
Twitter is one of the most popular social networking sites in the world,in which users compose their reactions in less than 140 characters .It has currently six million unique visitors every month and 55 million monthly visits.It often acts as a platform where celebrities interact with their fans ,hence it's importance in shaping a public opinion cannot be ignored. Also ,Twitter is used by music labels for the social media marketing of their music artists. In this project ,we attempt to explore behaviors of music on Twitter ,particularly music artists-how music tastes are reflected by Twitter and understanding the demographics of the music artists.For this purpose, we used data supplied by Echonest API -the backbone for the popular "Million Song Dataset". We attempted to build a recommender on the basis of data collected using content-based approach.
In the below map, we attempt to visualise locations of some of the most popular socially active and trending music artists in the world.
Who are the 10 most popular artists ? On the basis of the 'hotttness' parameter value which shows how much an artist is trending currently on Twitter Out of 1000 artists,we narrowed down 10 most popular artists - highest popularity obtained by Walking on Cars ,followed by the crowd favourites such as Taylor Swift,Eminem,Calvin Harris. Hotttness ranged between 0.99-0.93.
On the basis of familiarity value for each artist(ranging between 0.38-0.90),we concluded that The Beatles was the most familiar band;followed by David Bowie,Coldplay,Rolling Stones and Eminem. We can also say that on observing their start years,the older the artist-more familiar he/she is.
It is noticed that majority of the current artists have low familiarity even though they are popular.