Appetite for Stats: Insights on the world's most dangerous band
Guns N' Roses is one of my two favorite bands (the other is The Ramones) and I just want to know a few things about the songs of the Live Era GnR (1987 - 1993) AKA the real Guns N' Roses.
What's the average song duration? Is Civil War a long song for GnR standards? What about Perfect Crime, it's too short? And the lyrics... which are the most used words? How many words in average do the songs have? How many F* words have been used? Which is the song with the highest lexical diversity?
To answer these questions I analyze 44 songs from Appetite for Destruction, GN'R Lies and Use Your Illusion I and II. All songs that weren't written by GnR where left out of this analysis with a total running time of 3 hours 36 minutes and 20 seconds.
Average song duration
So, how long in average does a GnR last? Well, about 4 minutes and 55 seconds (mean) with a standard deviation of about 1 minute and 55 seconds. This means that a regular GnR songs may last something between 3:00 minutes and 6:50 minutes.
So, Civil War with a duration of 7:42 minutes it's actually a long song and Perfect Crime with 2:23 it's a short song.
Here's a box plot with the duration of the songs. You can see that with a duration of 10:13 minutes Coma is the longest song in the GnR discography.
Words
And what about the words? How many? How many different words? How many words in average? Here's the data:
But what's the meaning of a 12% lexical diversity? Well that means that 1 out of 8 words is unique. How do you calculate it? Easy: Unique words / Total words. This also means that there are about 39 unique words per song.
The song with the highest lexical diversity is Coma with 2,2%.
And which is the longest words used: sociopsychotic, 14 characters long. It appears in My World.
Now lets talk about the most used words. The two following cloud charts show the 30 most used words that have more that 4 characters long and 5 characters long.
Most used words larger than 3 characters
Most used words larger than 5 characters
One last thing here... what's the distribution of word length?
The F* word and related words
The F* words and it variations (motherf*) appear 42 times, that means that they only represent the 0,3% of all the words.
Data and tools
You can dowload the songs duration ods from here.
The box plot and the barchart were genereated using this R script.
The tag clouds where generated with TagCrowd.
The songs analysis was made using this Python script that uses the fantastic NLTK natural language processing library. You can get the songs list from here.