Data Web Scraping: American Music Awards
The skills we demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
And the American Music Award Goes To...
Intro:
The American Music Awards pays tribute to today’s most influential and iconic artists. The show, created by Dick Clark, first aired in 1974 for ABC, once the network's contract to air the Grammy Awards expired and moved to CBS. Data stats shows that the AMA's is now one of the biggest music shows in the world and is seen in more than 160 countries.
While the Grammy Awards (org. 1958) are awarded based on votes by members of the National Academy of Recording Arts and Sciences, the AMAs are determined by a poll of music buyers and the public. Nominees are based on key fan interactions as reflected in Billboard Magazine and on Billboard.com, including album and digital song sales, radio airplay, streaming, social activity and touring. These measurements are tracked by Billboard and its data partners.
So many award shows! Why American Music Awards?
This past year's American Music Awards aired in Los Angeles on October 9, 2018, on ABC. I vividly remember news breaking that Taylor Swift now holds the record for most AMA wins by a female, breaking Whitney Houston's record of 21 wins. I was then curious about other record wins over time and also how the categories have changed over its 44-year history. So I decided to scrape all 44 pages (years) of their winners' list to gather data for analysis.
Data Web Scraping:
I decided to use Python's Selenium Webdriver library for scraping because of the simplicity of the website's design; the URL's are easily looped by using a numeric range from 1974-2018. The main data was also easily scraped from a simple results table. Each web page scraped contained a results table with 3 columns for each year 1974-2018. After the script finished running, I had a table with over 1000 rows of every winner in the American Music Award's history.
The Data:
The original columns scraped were Year, Category, Winner. Using a python jupyter notebook, I further manipulated the dataset to gather more info. First, I split the original 'Winner' column into 'Winning_Artist' and 'Winning_Album'. To get a better understanding of what categories (album, artist, gender) are awarded, I extracted certain keywords from the original 'Category' column and made 2 new columns 'Cat_Type' and 'GenderAwards'. Finally, I created 5 new columns to count 'Category', 'Cat_Type', 'Winner', 'Artist_Wins', and 'Album_Wins'.
The Data Analysis:
While some findings were expected, some were quite shocking. To warm up, let's take a look at the top winning artists:
To see the most awarded artists, I plotted a word cloud of artists who have 5 or more wins.
- Michael Jackson, Taylor Swift, Whitney Houston, and Alabama are the biggest winners
Distribution of Total Artist Wins > 5:
This graph answered a lot of questions for me. To plot the top artists, I filtered the dataset for total artist wins > 4 and sorted from highest to lowest total wins (Michael Jackson at #1).
- First, I wanted to know who was the first artist to win consistently over a long period of time? The AMA's debuted in 1974, from its first ceremony and for the next 2 decades, Stevie Wonder looks to be the first consistent winner.
- Any difference in the top 3 artists across decades? It's also worthy to note that 2010s are the first decade with the top 3 artists to be all female (Taylor Swift, Rihanna, Carrie Underwood). Before that, Michael Jackson and Alabama dominated the prior 3 decades.
- Which is more common, winning big for a short period of time or over many years? Winning consistently is definitely the most common trend among top artists.
Yearly Artist Wins > 3:
This plot demonstrates the total wins for an artist on a yearly basis. The idea here is to investigate if an artist ever goes on a winning 'streak' during specific years/time periods. The image above of Whitney Houston with 8 trophies in a single year was striking to me. I wondered how often that happens or if had ever been done before.
- How often do artists get a record number of wins in a ceremony? Just 2 years after the AMAs inception in 1975, the bar was set quite high to 4 total wins by Glady's Knight & The Pips, they never won big again after that year. However, 9 years later, the next record was broken by Michael Jackson with 8 wins, doubling the previous record! This record was matched once by Whitney Houston 10 years later in 1994. They both currently hold the record for most wins in a single ceremony.
- Any slow periods with no 'mega' winners? It's notable in 1995 -2008 there was no artist even close to 3 wins in a ceremony. Taylor Swift changed that in 2009 with 5 total wins that year.
- Although Taylor Swift hasn't broken Whitney and Michael's iconic 8 wins, it's clear to see that she's the only consistent big winner of the past 2o years.
- Has anyone come close to breaking Michael and Whitney's record? 23 years after Whitney won 8 awards in the 1994 ceremony, Bruno Mars wins 7 awards in the 2017 ceremony. He is the only artist to come close to her and Jackson's record.
Gender-Related Categories:
Note: While analyzing this graph I found insight that has nothing to do with gender. It is a sad moment in American history, but the data also shows how much culture and musical influence impacts these categories over time.
- Are there any consistent gender categories over time? Country, Pop/Rock, Soul/R&B are the 3 most consistent gender-related genre categories over time. All other categories (besides Rap/Hip-Hop and Disco) have never been assigned a gender.
- It's also worthy to note that these 3 categories had a name change in 2016 and have carried on under that name since.
- Any inconsistencies to these gender categories? In 1985, the Soul/R&B Female and Male categories were replaced with Black Male/Female categories... it was never repeated after that year. I did a bit of research on that year's ceremony but couldn't find any explanation for the category change.
- It's also interesting to note that Rap/Hip-Hop Female category was only around in 2005, the solo Male category continued on until 2010.
Distribution of Category Types:
Next, I wanted to understand the consistency of what awards are given out over the years.
- What categories are the most consistent? Album, Male/Female Artist, and Group are the only category types that have been around since the ceremonies inception.
- Were any of today's categories created at a later time period than I expected? Looking at 2008, I found it shocking that Artist of the Year was created.
- Also shocking that Soundtrack wasn't a category until 2010. Perhaps they were combined with regular album categories beforehand?
- Any name changing? It looks like Single of the Year was more or less replaced by Artist awards with no gender. And then in 2013, it made a comeback and was then renamed 'Favorite Song' which is now current.
Album Wins > 1:
- Looks like the top album with 4 wins was somewhat of a duplicate because it was awarded as a duet and album.
- Camila Cabello's album, however, in the 2018 ceremony was the first album with 3 wins since Whitney Houston's album in the 1994 ceremony.. 24 years!
- The 2000s are the only decade lacking an album with 3 wins