By Anna Leah
In 2013, a new hashtag arose in the wake of – #BlackLivesMatter. The following year, in Ferguson, Missouri, and on Staten Island.
The public reacted with protests and conversations about racism in America that stretched . During the first tumultuous years of the BLM movement, upheaval was not just felt in activist circles and halls of power ��� hip-hop explored new topics as artists reckoned with the moment.
���Hip-hop lyrics isn���t just about the music, it���s a living record of American history,��� says recent York College graduate . Some artists rapped about ���racial injustice, identity, resistance, while the other half of the group is talking about lifestyle, money and stuff like that.���
With Devonte Edward, Daniels examined how the words used in popular hip-hop songs mirrored each other and aligned with major world events. Both studied computer science, so they approached the task as programmers as well as music-lovers.
Their first challenge was the lack of existing data sets. Creating something testable took up a majority of their time. Their mentor, , encouraged the team to define and categorize an experimental body and then find a question they could answer before graduation.
The research pair accessed a database of Billboard 100���s chart-topping songs since 1990. This excluded independent hip-hop from their study, but it was a data set that the undergraduate team could use. Billboard���s list gave them the artist, title, and ranking of a hit, but not its genre.
They then used three different AI programs ��� OpenAI, Grok, and Gemini ��� to code each song with a single genre. Daniels explains that they used different programs to balance each application���s blind spots, achieving an accuracy of over 90%.
They filtered almost 3,000 songs into about 800 hip-hop tracks. They then assigned a vector value to each piece based on an automated analysis of the lyrics ��� which limited the weight of implied meaning, slang, and . They compared those values to those of other songs year by year to see lyrics��� diversity and how that landscape changed.
The team found a genre finding itself. The 1990s were fairly scattered in topic. Around the same time as the loss of pioneering artists like The Notorious B.I.G. and Tupac, the lyrics started to move towards each other.
Edward says, ���As time went on, it came together and it was all similar.���
Those similarities peaked in 2004. These coalescences occur in cycles of roughly a decade. We are just coming out of a slight dip in diversity that occurred in 2025.
Daniels says that changes in hip-hop lyrics reflect shifts in technology as well as culture. She is curious about how future scholars will track the adoption of artificial intelligence ��� which could lead to fewer individual fingerprints in the song���s poetry.
The project���s mentor describes undergraduate research as a process. Daniels spent her fall term preparing a paper about weather prediction. When the time came to expand the scope and findings to be published, Srivatanakul encouraged the students to follow topics that interested them enough to inspire important questions. ���When you put more time into it, I think, it���s not about just coming up with the research,��� she says, ���but coming up with the research that you���re passionate in doing.���
To create a publishable piece from their hip-hop database, the team will explore keywords to see how they align with lyrical themes throughout the years.
Working on this project also gave Daniels a chance to play some of her classic favorites. And what she learned impacts how she hears the songs now.
���Before, I���m just listening to it,��� she says. ���What it made me realize was that what is happening in the world and how it���s impacting the artists themselves really reflect in the lyrics.���
