K-Pop Space addresses the needs of dancers who want to create covers of their favorite K-Pop songs by allowing them to sort through YouTube videos based on tags such as number of dancers, gender, and entertainment company.
Introduction: K-Pop Dance Music & Its Globalization
The spread of Korean entertainment and culture, known as the “Korean Wave,” includes K-Pop, or Korean pop music, which has grown from a genre into a subculture among teenagers and young adults all over the world. The driving forces behind K-Pop as a global phenomenon not only includes catchy musical elements, but also dance routines and visuals such as props and colorful costumes. These characteristics of K-Pop contribute to its popularity because audiences do not necessarily have to understand Korean to enjoy the performances, as dance is a form of language that transcends cultures. Another catalyst to the globalization of K-Pop is the growth in digital media and social networks such as YouTube, not only expanding the reach of the music videos, but encouraging the creation of online communities dedicated to the genre, including those who create dance video covers to share on this platform.
We are a group of students and faculty members from the Digital Humanities program at UCLA. For this project, we built upon previous work of faculty who created neural network animations of top K-Pop dance videos from YouTube. From there, we had to decide what route we were going to take this project. We narrowed down the massive group of K-Pop videos we had on file to the top 300 videos, cross-referencing our data with K-Pop Database. We then gathered metadata on each video, including artist name, entertainment company, number of members, and more.
Our next step was to decide how we wanted to present our data, and we decided on a WordPress website, with the target audience being people who wanted to create K-Pop dance covers. Our search engine allows users to navigate the site and narrow down what type of video they want to look for. In addition to our search engine, we included data visualizations based on algorithm that identify how many bodies are in each frame at certain times, as well as average the body count.
Because of our limited timeframe to create the website, we have some limitations we hope to address in the future such as video search functionality and overall design, as well as take a deeper look into applying AI, Deep Machine Learning, and Pose Detection to our research.
We hope you enjoy!