A visualization on the use of language in Hip Hop, created by Yodahe Alemu and Joshua Verdejo
Our rationale for our A4 submission which our final project was built on top of.
Similar to our work breakdown for A4, we split the work such that Yodahe focused on the website while Josh focused on the datasets.
Joshua primarily handled the creation of our data sets. This included finding current online databases we could pull from, working with different web APIs to build our own unique dataset, parsing through any unnecessary data, and cleaning up the data for use with the website. Joshua also helped with the design of our visualization and it's main interactions, as well as doing final touches on the actual website.
Yodahe primarily handled the development of the website visualizaiton. This included working with d3.js, building out the main interaction systems, creating the visualizations, and connecting the datasets to the website. Yodahe also helped with the design of our datasets and doing research of current online databses, as well as doing some touch ups to the datasets for the purposes of connecting them to the website.
This version of Hip Hop Talks, which was built on our initial two weeks of work in A4, took a month of work to build.
In the beginning, we started with ideation and how we wanted to expand upon the ideas set out in A4. In A4, we had tackled the very simple question of "What words do Hip Hop artists say most?". In our final version, we decided to expand on that question by providing more artist-specific insight into our datasets. We also wanted to build on the playful experience of our original design (which came from typing out words) by including a "next word suggestion" feature. This feature would tell users the most likely next word to be said by an artist based on the word the user had typed. We had originally wanted to include this feature by syncing with the Spotify API and giving users a word suggestion based on their favorite artists; however, we quickly realized that the processing of this data would make it nigh impossible to have a fluid visualization/interaction, so we opted for simply using the artists already in our dataset.
Once we had the idea solidified and had found some online data sources we could scrape from, we split up so that one of us could focus on the construction of our dataset (we needed unique datasets since nothing that existed online exactly fit our needs) while the other built out the framework for the new visualizations that we would need. Even though we worked on separate tasks, the entire process was very much collaborative as we met often in order to discuss progress that had been made and what next steps should be taken. Because we split up the work in this way, the development process was very seamless. Once the datasets we're finished, we were almost immediately able to finish the website since the website had already been built up. Since there was a good amount of communication, the data that was expected by the website exactly fit the dataset that had been prepared.
In the end, we were able to build out a visualization that gives users an insight into the language utilized in the Hip Hop industry, both on an industry-wide level and an artist-specific level. We were also able to create a fun and engaging interaction system around our main typing module as well as different clickable parts of the visualization.