>> From the Library of Congress in Washington, DC. >> Next we have Tahir Hemphill, an award-winning multimedia artist working in the areas of interdisciplinary collaboration, thought, and research. He's planned business strategy in the entertainment, advertising and non-profit industries. He also produces art and media arts education programs and manages the Rap Research Lab. He's going to share a story about how visualizing rap lyrics helps to teach data literacy to high school students. Tahir, you're up. >> Tahir Hemphill: My name is Tahir Hemphill. I am a artist, educator, radical archivist, and a creative technologist. And I'm going to talk about my projects at the Rap Research Lab. My practice investigates the role that systems play in a generational form, and also the role that collaborative knowledge production plays in the resilience of communities. So the slides that I'm going to go through today will fit in those two areas. So the system that I developed -- the one system that I developed is the Rap Almanac. And it's an online searchable database, ^M00:01:19 ^M00:01:22 and it contains the natural language processing of tons of hip-hop songs. Right now, there are about 25, 30 artists, about a quarter of a million songs, and this is some of the data analytics associated with the songs. So in a database, we have artist name, song title, the album title, geo location -- everything is geo-tagged -- and has release dates for all of the songs. And some of the syntax analysis is the word count, reading level, rhyme density, and this rhyme analysis that will tell you if what you're reading is a perfect rhyme. And also it tells you where the rhymes fit within the lines. And the sentiment analysis covers sentiment, keywords, people, and places. And this is what a search in the database looks like. In this case, the user typed in "power." And then you get back this wall of data -- song title, album title. And then you can add these advanced search options. In this case, they're adding sentiment and people, syllables per word, and redoing the search. And he most exciting part about this, well, besides the YouTube view is the fact that you can press a button and get a machine-readable output and a human readable output. So you can export to CSV and export to JSON. And it took a long time to get to that point. It seems really small but I was ecstatic when we got to that point. So, what the database gives us is a new context for hip-hop, right. So there's content, there's a location tagged to the content, and there's a time. So for every utterance in hip-hop there's a time and a place on earth. You can do, in that example, the search for "power," and find every rapper that rapped about power, where in the world they rapped about power, when they rapped about power, and the context. So it's a big data rap project. So I'm going to go through a few slides of examples of visualization. So one of the first things I did immediately was to just to dump all of this in Google Earth. And what you're looking at is, or what you just saw was Staten Island rappers. And now you're looking at Brooklyn rappers, and then Queens rappers. So the project maps the geography of the language of hip-hop. So you can find out how slang travels, where slang originates, how it travels, how each region influences other regions. And that can be done for slang, that can be done for politics, that can be done for many other things. Another data vis, early data vis that I worked on, looked at champagne-brand mentions in hip-hop from 1980 to 2010. And of course, it's like it's quaint, it's funny. Like, you know, there's -- we all have an image of rappers and champagne use, right? But if you also look at -- that image that we get from music videos. But if you also look at champagne as a aspirational product and times for champagne use that are tied to expressions of celebration or expressions of wealth or the fact that you want to pretend like you have wealth, then you can dig deeper into the cultural meanings of champagne and rap. This is a network graph that I made that looks at contemporary and modern painters that are mentioned in rap. And Jean-Michel Basquiat became really popular recently and it's called Picasso Baby. And again it shows these connections, this different context between rap lyrics and culture. And a recent -- well, a project that I made last year is called Spotify City, and it deals with virtuality. So I had a residency at Spotify, and what I did was look at the content of the streams as a representation of a listener's aspirations and desires as revealed through the music that they listen to. So the idea is that if you're driving to work, on your way to work, you're probably not listening to song about being at work. You're listening to a song about being somewhere else at work, and how do we extract that virtuality from streams? This is a light pen drawing made from a 20-foot-tall industrial robot. So using the semantic analysis, I extracted the locations that are mentioned in 12 rappers' bodies of work, from their first song to their last song, and through -- I think it was like a 14-step process, sent these codes to a robot arm and put a light pen in a robot arm's hand and had the robot arm trace the locations around the globe. In this case, it's Kanye West. This is a digital model of the same thing that we are developing into a VR educational piece. So the green paths are the paths that the rappers -- the green paths are the paths that the rap travels. ^M00:07:51 ^M00:07:58 And I wasn't satisfied with just two-dimensional -- or I wasn't satisfied with two-dimensional representations, so I'm also producing data sculpture of this work. And this is the suite of all 12 rappers. And a lot of this work is going to be exhibited. I'm having a solo show at CCA in September. Everyone here is invited too. And I'm going to present this work. So the project visualizes hip-hop lyrics as a cultural indicator. I mean, you can imagine overlaying data sets, health data sets, crime statistic data sets, education data sets down to a county level, and overlaying those with this data set. And this is data prahn [phonetic]. This is the analysis of -- the semantic analysis of Drake's entire body of work being processed. ^M00:08:59 ^M00:09:05 Cool, but data needs memory, right? It's nice to have -- it's nice to have pages and pages of data, you know, whether it's machine readable or human readable, but in order to tell a story with data, you kind of need an experience tied to it to bring it to life. And it was this question or these types of questions that made me start the Rap Research Lab. And especially like this question: should a culture own its own data? Like, who is responsible for protecting this data? Who is responsible for telling the stories about this data? Who is responsible for checking the stories told about this cultural data? And even when technological development advances its workers into exciting new spaces of creative inquiry, the opportunity to challenge traditional modes of hierarchy is often missed. And most of the time, if there's a new space open in science, we bring our same biases to these open spaces; instead of trying to do something new or imagining something new, we just [inaudible] our old self to the new spaces. So in the age of big data, normative narratives around hip-hop reproduce stereotypes. And also that -- for me that also extends to the singularity and artificial intelligence. Like, you know, the singularity in a way is a colonizing process. It ignores what happens on the sides and just focuses on what works for the system. So I founded the Rap Research Lab to have a space and to invite people to talk about these questions, explore these questions and grapple with these questions. And the Rap Research Lab is a creative technology studio, and we explore rap as a cultural indicator through educational, editorial, and creative interrogations. I'm going to read a bit. We are a team of researchers, radical educators, technologists, and artists who use rap as text through which to better understand popular and youth culture. We explore questions of social justice and facilitate resilience through community data projects. And we design projects and ask questions both collaboratively and individually that challenge notions of ownership and agency in these spaces. So just a few axioms, rules off the top that we believe at the Rap Research Lab. One is rap is the most influential global and youth art form, period. According to a Spotify analysis of 20 billion streamed songs back in 2015, rap is the most popular, most listened to genre in the world. According to a recent study, rap is more popular than rock, but we already knew that. I mean, rap has more words. Another axiom is hip-hop isn't rap, graffiti, deejaying and breakdancing; hip-hop created rap, graffiti, deejaying and breakdancing. So in the beginning there was a deejay. There were no hip-hop records. So Kool Herc playing in 1973 in the Bronx, you know, that's cited as the first hip-hop party. There were no rap records. There were no hip-hop records. There were soul records, there were funk records, there were blues records, there were rock records. But what was hip-hop was the way he played those records. So hip-hop is a framework, and the framework in which he played those records. And this is a quote from my colleague David Goldberg. This framework is helpful. In a fragmented world, hip-hop is an effective approach to problem-solving, especially when your future depends on rising from the ruins of inherited mistakes. So because of this framework, we decided to teach. And this is a map of our curriculum and our lesson plans. ^M00:13:30 ^M00:13:34 And we start with an introduction, you know, the history of the technology, the history of the culture. And then we teach the students how to roadmap their research. So we teach them how to ask questions, how to formulate good research questions. And then we have a content coding exercise where we do closed readings of the lyrics and content code them. This is analog. We also have a digital version. And then we teach the students how to conduct a search, how to use the tools. And then we deal with data bias, like their own implicit bias and then external bias. So how did they collect the data? What did they leave out? What was left out in the initial process of the data that they collected? And how things fall off along the way. And then we have some mapping lessons and visualization lessons. So I'm going to read it again. So the methodologies in this framework, we invite participants to engage critically with cultural research, data collection, data ethics, implicit bias, data bias, and storytelling, right? And these are the elements that ultimately empower the critical thinking and decision making that leads a citizen to become more capable of establishing informed positions. This outcome not only influences behavioral change but is a precursor to developing and deploying a healthy democracy, especially in this post-truth era. So we're dealing with data literacy, media literacy, and actually listening to rap and having fun doing it. I'm making this sound so boring, but it's really cool. ^M00:15:34 ^M00:15:37 So this is -- so the next three slides, so in terms of development and process, we piloted a few years ago, then we scaled up. I invited teachers to produce the curriculum with me and then a third phase was the spread. So this is my one-room school house in the Bronx, South Bronx. And down on the bottom right, this is a design sprint that we have where we invited educators in to test our curriculum. And on the top right is actually the graduation day where I invited some of my colleagues in the art and tech world to come and give critique and support to the students. Top left, that's a picture of me waxing poetic. And on the bottom left is one of our content coding exercises where we ask the students to identify in this case five themes that they think they'll find in the music. In this case, it was body sexuality, and then achievements, and then relationships, love, family, and then conflict, and origins, politics. And then we give them text and then we ask them to content code. So every time they see one of these themes, they tag it in the text. And that's the analog process that helps build algorithmic thinking. And then it's also modeled in a digital process. And this is the scale. This is when we scaled up. And these are some of the workshops we did in collaboration with the Parks Department, Eye Beam and Hive, Mozilla and Creative Capital. And then the last phase that we just finished last year was the spread where we took our curriculum and we trained New York City Parks Department educators to implement our curriculum in ten locations, in five boroughs, for nine months around New York City. And it was a trip. It was a trip to see students study me and my process on their screen. But it was exciting to see that this process could be spread and scaled so I wouldn't have to be the only one standing in front of the classroom. And these are some of the -- I'm going to look at two, show you two student projects. So we have a student research page. And one student, Gabriel, he was interested in doing research on the N word in hip-hop, in all of hip-hop. And he quickly realized that that was too much data for him to process, right? So he was kind of bummed out, and I was like, "So what do you want to do?" Then he realized in his research that there was one artist that had a song where he mentions the N word like almost 200 times. So he decides to focus on that artist and content code all of his songs through a lens of this dictionary of slang that gives seven or eight -- that gives eight meanings of the N word. So what you're seeing is a digital content coding project from Gabriel. So these are the songs on his dropdown menu, and on the right, these are the different descriptions. So a close friend, the black experience, just a generic term, someone's lover, someone that's rebellious, something that's disrespectful, something that's negative, and also just a filler, just to, you know, to fill sonic space. This is the project by Emmanuel who was reading some of the articles about rappers and raps that would be used as testimony in court. And he wanted to investigate the differences, or whether or not there was a difference between crimes that were rapped about and actually crimes that were happening. So this is a map that he made. And you can see that there's -- I mean, correlation and cause aren't, you know, aren't tied, but there's a difference. So when crime is down basically crime raps are high. And I'm going to close with this one quote by DJ Premier, and he's talking about producing Jay-Z's Reasonable Doubt album, his first album. And he says basically one thing he loves about hip-hop is that you have to understand how to listen to it. Just because Jay-Z is speaking in English doesn't mean you understand what he's saying. And that's the reason for the Rap Research Lab, to do a deep dive into and decoding the most popular art form in the world. Thank you. ^M00:20:50 [ Applause ] ^M00:20:54 >> This has been a presentation of the Library of Congress. Visit us at loc.gov. ^E00:20:59