Sentiment Ananlysis of Billboard Hot 100 Songs over Time (1958-2019)

For this project, I scraped all song lyrics for all entries of the Billboard Hot 100 from 1958 to 2019 from the website genius.com using LyricsGenius, a python API client developed by John W Miller. The dataset consists of lyrics for 26,138 unique songs. After scraping all the song lyrics, I performed sentiment analysis of songs using TextBlob, a Python text processing library.

Results

1. On average, the sentiment of lyrics on the Billboard Hot 100 tend to be neutral.

2. Lyric Setiment in 2019 is about 4 times more negative as lyric sentiment in the 1960s.

3. Lyric sentiment, on average, has gotten 1.3% more negative annually between 1958 and 2019.

4. Top keywords for lyrics in 1958 include:

   1. "like"
   2. "come"
   3. "littleness"
   4. "manning"
   5. "knows"
   6. "jump like"
   7. "good"
   8. "time"

5. Top Keywords for lyrics in 2019 include:

   1. "like"
   2. "yeah"
   3. "niggas"
   4. "bitches"
   5. "lil bitch"
   6. "love"
   7. "need"
   8. "fuck"

Data

The datasets containing the analysis results and scraped lyrics could not be uploaded to github due to size limits. It can be viewed and downloaded here: https://data.world/szubair/lyric-sentiment-ananlysis

Author

Salim Zubair