Songs by the famous British rock bands ``The Beatles'' and ``Queen'' are still being played in various places. Many people probably hum or sing this song every time they hear it in a restaurant, on a TV commercial, or at a sports festival.
Their songs may have simple lyrics, but they may represent the emotions they express in their lives. You can't say things like "You can't buy love!" or "We're champions!" in everyday conversation, but you can easily quote the lyrics of a song.
In this blog series (scheduled for 3 parts in total), we will analyze the emotions contained in the lyrics of "The Beatles" and "Queen" and the differences in the words used by each band.
In this Part 1, we will perform a sentiment analysis on the lyrics of "The Beatles."
In the current latest version of JMP Pro 16, Sentiment Analysis* is available as an option in the text analysis platform (Text Explorer), so let's use it for analysis.
The analysis target is the lyrics of 171 songs by The Beatles, excluding cover songs, songs in languages other than English, and songs released after the band disbanded.
*Sentiment analysis is a JMP Pro only feature. This cannot be done with JMP. Furthermore, the only language in which sentiment analysis can be used is English. It is not possible to analyze Japanese.
Sentiment Analysis Report (Sentiment Summary)
Sentiment analysis searches for positive and negative words in a document, and provides a score to indicate how much emotion the sentence is filled with. For example, in a survey, when customers leave open comments about their impressions of a certain product, it is used as a method to analyze whether those comments are positive or negative, and is also known as positive-negative analysis.
The figure below is a summary of the results of emotional analysis of the target lyrics.
First, be aware of the colors in your report. Those painted in an orange-ish color mean positive, and those painted in a blue-ish color mean negative.
In the report on the top left, the emotional score is calculated for each song. Here, Document = 56 is selected, but let's take a look at the lyrics of "Hey, Jude" on line 56 of the data table I created.
In the bottom left, the lyrics to "Hey, Jude" are shown, with the positive and negative words colored in. Additionally, those words are scored on a word-by-word basis*. The score table is shown on the top right.
This calculates the sentiment score for "Hey, Jude" to be 38 points in the Overall Score column. This is calculated using the following formula.
Overall Score = (Positive Sum(sum of positive word scores) + Negative Sum(sum of negative word scores))÷ Number of words recognized as emotion words
For "Hey, Jude", the total positive word score is 520, the total negative word score is -100, and the number of words recognized as emotion words (colored) is 11, so the emotion score is , (520 - 100) / 11 = 38 points.
Similarly, emotional scores are calculated for other songs, and the distribution of emotional scores is shown in a histogram in the upper right corner.
* The words that are considered positive/negative and their scores are determined by default in JMP Pro. Analysts can also add words and change scores.
Are the lyrics positive? Negative?
However, in this analysis of lyrics, since lyrics range from short (minimum number of words = 17) to long (maximum number of words = 293), we will divide the emotional score by the number of words in the lyrics. Use the following Overall Score by Word Count (emotion score per word count).
Overall Score by Word Count = (Positive Sum + Negative Sum) ÷ Number of words
The histogram and summary statistics of Overall Score by Word Count calculated in this way are shown below. Here, the N number is 123 because we are only targeting songs that have some kind of positive or negative emotion.
Looking at the histogram, there are more positive songs than negative songs, with an average score of 1.67. Overall, you can see that more positive words are used than negative words.
What are the emotions in Lennon's and McCartney's poems?
It is said that many of the songs by The Beatles were written by Paul McCartney and John Lennon. Therefore, let's compare the emotional score "Overall Score by Word Count" calculated in the previous section for each song written by the two.
(Some songs are jointly written by Lennon and McCartney, but information on the Internet or other sources may reveal who wrote the lyrics. In that case, they should be included in either the Lennon or McCartney category.) )
The figure below is a violin plot showing the distribution of scores when the lyricists are divided into Lennon and McCartney.
From the figure, it can be seen that McCartney has fewer negative songs and some positive songs compared to Lennon. On the other hand, Lennon has one very negative song.
Let's examine the emotions of McCartney's songs with high scores, as shown by the labels in the diagram above. It greatly contributes to the words "better" and "love".
Meanwhile, let's examine the emotion of Lennon's lowest-scoring song. There were some negative songs like this.
How do you like it? As you know, lyrics generally express emotions by using the same words or refrains over and over again. Therefore, songs that repeatedly use emotional expressions with high scores will have high or low scores.
However, I think that the fact that you use it so many times is also a sign that you want to express your emotions fairly strongly.
As the famous Japanese song goes, "the refrain is screaming."
In the next blog, we will analyze the sentiment of "Queen" and compare the emotional expressions of the two bands.
by Naohiro Masukawa, JMP Japan
