Data visulization
Figures about collected data and discussion on the findings.
Artists data set
For Artists data set, we analysis the correlation between popularity and follower. The scatter plot shows that they are relatively positive correlated. Also we sum up the followers in various genres and pick out the top 20 popular genres. They are shown in the figure below.
Songs data set
Spotify provides 9 main features to describe songs so we plot the histograms of these attributes.
As the figure shows, the 'acousticness', ' instrumentalness', 'liveness', 'speechiness' are left-skewed. 'energy' and 'loudness' are right-skewed. Positive songs are less than negative songs exhibited by 'valence' and more songs are energetic.
Also we use a heatmap to find the correlations between each couple of these features. Only few features are correlated with each other, 'Loudness' and 'energy' and 'energy' and 'acousticness'. Each feature is different so we can more accurately describe a particular song.
We choose 4 well-known genres and 5 common features to find their differences.
The common trend is that the louder the more energetic. They are positively correlated. Except these two features, there is no significant correlations, which proves the finding before. Pop music is louder than other three genres. It and blues music has less instrumentalness but classical and blues music needs more instruments. Classical music has a large amount of music with lower loudness.
You can control this graph to explore more interesting findings!
Influence data set
As mentioned in the data cleaning section, we add a new feature gap_year in influence data. Thus, we use a histogram to show the gap year between influencers and followers.
Most years are gathered in the range from 0 to 30. It shows that most musicians are influenced by their peers and those in the last generation. Very few people are inspired by those before half of the century. It also should be noticed that some gap years are smaller than 0, which suggests some musicians are keeping in progress and following the contemporary mainstream.
We also draw the bar plot about gap year based on influencer's and follower's genres.
Influencer's years
Follower's years
Influencer's years
In terms of influencers, Classical and Vocal are the longest genres while Electronic is the shortest one. When it comes to followers, Easy listening is the shortest while Classical is the longest as well. It implies that Classical music can influence longer and those with shorter gap years may change styles frequently.