Histogram
In the left histogram, we plot the distribution of popularity across years. We equally divide movie years are into three groups, each have equal range of 10 years. Form the plot, we can observe that the popular movie tracks with popularity score higher than 60 are distributed equally among the three bin groups. However, the unpopular tracks with score lower than 20 are concentrated in year 1989-2008, especially for the tracks with score less than 10. The distribution is highly imbalanced.
In the right histogram, we plot the distribution of popularity with tracks has low instrumental feather and tracks with higher instrumental feature. It is believed that a track with high instrumental are songs without human voice, while track with low instrumental usually has lyrics and human voice. From the plot, songs with low instrumental feature has significantly higher popularity score than the songs with low instrumental feature, which indicate that the public usually prefer songs with lyrics than the pure instrumental songs.
In the left histogram, we plot the distribution of the count of low instrumental track in each movie. We considered the instrumental value as the track with instrumentalness lower than 0.05 confidence level. And it turns out the songs with instrumentalness lower than 0.05 are usually songs with human voice. From the distribution, we can observe that over 20% of the movie does not included a song with lyrics.
In the right histogram, we plot the distribution of the total count of the tracks in each album. From the plot, we can observe that there are nearly 15% percent of the movie only have one songs in its album, it is perhaps because that Spotify does not includes all the soundtracks of the given movie in the albums. Besides that, most movie has song track with in range 10 to 20. Thus, a number of 10 to 20 songs would be the most comment soundtrack to fill the requirement of its plots.
Scatter Plot
Box Plot
Tree Map
This tree map can directly show the movie gross components in different years. Each area represents a different genre of movies. The larger the area is, the more gross movie in this genre got.
Bar Plot
This bar plot exhibits the number of movies in different genres. The scale in on a 10-year gap. For movies covering more than one genre, they will be counted several times.
Line Plot
In our analysis, we selected the Harry Potter Series as an example. From the first two charts based on <The Prisoner of Azkaban> and <The Goblet of Fire>, we can see the highest peaks both reach more than 0.6; in contrast, on the latter two charts based on <The Order of the Phoenix> and <The Deathly Hallows>, both peaks of valence are less than 0.4. These results are consistent with the actual plots of the movies, as earlier series were much more exciting, while the later series were relatively more depressed and gloomy.
Bubble Plot Show Video
This is an interactive bubble plot with x-axis as the movie rating and y-axis as movie run time. We divide the common movie genres into 13 groups, each group has a combination of 2-3 genres. The size of the bubble indicates the number of movies with that genre group. The more movies of that genre group of a given year, the larger the bubble is. The slider at the bottom of the graph enables us to slide on the years that we are interested in.
By looking at this interactive bubble graph, we have two major findings. The first is that there are very small portion of the movies have high ratings. By moving the slides, there are only a small amount of movies that have ratings higher than 8.5 The second finding is that high rating does not necessarily lead to high rating. Some movies might be adapted from famous novels and thus receive a lot of attentions. However, the movie itself is not as good as expected even though a lot of people go to the cinema to watch it. This would explain the reason.
Radar Plot Show Video
We use the interactive radar graph to investigate the six music features of a piece of music or to make comparisons with another soundtrack in the same movie. We try to find some trend of the soundtracks within the same album. The music features that we look at are: acousticness, instrumentalness, energy, danceability, liveness and speechiness. After looking at different movies’ soundtracks, we could not find any determined trend because different soundtracks turn out to vary a lot from one another. In the following graphs, we randomly choose four soundtracks from La La Land. In the left graph, we compare the ”Planetarium” VS “Another Day of Sun”. They are not similar considering the degree of six music features. In contrast, in the right graph, when we compare the “Engagement Party” VS “City Of Stars”, there are a lot more overlaps. Therefore, we could not find any trend of the soundtracks within the same movie based on our observations on the radar chart.