This is an ongoing project that uses a corpus of hip hop lyrics to find interesting links between hip hop artists. By finding unique phrases that are only shared by a few artist, we can establish connections between them. These shared phrases could come about as a result of paying homage, intertextual allusion, or simple plagiarism. Outside the realm of modern intellectual property, this is a common an well accepted practice in blues music, a genre where rap music has deep roots.
I listen to a lot of rap music, and I've noticed this popping up. It seemed like a fun way to jump into NLP. Really, something about juxtaposing AAVE and technical/academic language just tickles me.
This project will show directionality of borrowed phrases, inferred by release date, and cluster artist around influential artists with the most sampled/borrowed phrases.
##Current state Due to the typographical and labeling inconsistencies of the corpus, the original idea to transform this into a network graph has been put on hold. For now I will be manually combing the matches to uncover interesting matches. I'll then weave a story around these matches and present them with context and links to the songs themselves. Check these out at http://www.noahlipsyc.com/personal/projects/hip-hop-web/
- Create a website that that allows the public to comb through the matches on their own. This site could present the user with a list of matches for them to explore. After selecting a lyric, it would then search for all the corresponding songs on YouTube, advance it to roughly the right place in the song, and present the lyrics bellow them.
- Find matching phrases between the hip hop corpus and influential literary works (Old Testament, New Testament, Koran, commonly quoted lines from literature)
- A network graph (maybe one day?)
Lyrics courtesy of the Original Hip Hop Lyrics Archive