Selection of Songs

Our research began by selecting 10 songs from 12 various hip-hop artists, these artists included:

In order to avoid bias in terms of selecting songs, we came up with a method. A random number generator was utilized, and from there with the numbers generated, we would select the 10 songs that corresponed with the numbers given across all of the artists' songs. These would then be the songs in which we would markup and anaylize. Songs and their lyrics were pulled from https://genius.com/ in order to avoid copyright. Eventually however, due to time constraints, we would end up having to drop three of these artists, these being MF DOOM, Earl Sweatshirt, and Travis Scott. We decided that this was helpful in the long run, as we were looking to research how artists evolve over their careers, and the ones we dropped had little to no variation overtime, therfore in the end dropping these artists would strengthen our results and produce ones that are more likely to be moreso vali

Markup: How We Did It

Each group member was given three artists with ten songs each, for 30 songs per group member, making 90 songs in all.

Tagging Tone and Mood

In order to mark up these songs and pull the essential information, we decided on tagging tone and mood for every intro, chorus, verse, bridge, interlude, and outro within a song. In order to maintain a sense of consistency throughout all 90 songs, the tone and mood attributes were signifigantly constrained. Mood could be tagged as any of the following: hopeful, energetic, cynical, apathetic, empathetic, determined, angry, sad, happy, confident, anxious, ambiguous. Tone was constrained to three: positive, negative, and neutral. In the insatance where a song had a featured artist that sang alone, which was quite frequent, the mood and tone attributes were set to "null," as we were not looking to analize anything that they sang or said because we were specifically looking for the tones and moods of the main artist. Tagging Topics Another essentail piece of information that we needed to isolate from the songs in order to help us track an artist's evolution over time was specifc topics that were being brought up throughout the lyrics. We decided on these such topics being: drugs, death, family, religion, money, poverty, wealth, violence, social issues, gangs, and ambiguous. Tagging topics would allow us to track how, as stated with our reseach question, an artists lyrical subject matter can evolve over time.

Process for SVG

The process for the bar chart svg was a bit more complicated than anticipated. At first we were going to have a chart for each song for each artist, but that would mean we would have to fit ninety charts on a page. To fix this, we decided to make one chart per artist, comparing the tones of their early and late albums. To do this, we had to use the collection function. Camryn worked on this with Dr. Birnbaum as that in and of itself was a difficult and complicated process. Essentially, the collection function creates nine collections, which contains the lyrics for each artist we looked at. When the template is applied to hip:artist, the template runs over each artist collectively, giving us our data.