During the last two weeks of every cohort, Launchers spend their days working on a capstone project, known as a "Breakable Toy." They are free to apply everything they've learned to build an app of their choosing, and most choose a project that ties into a hobby or cause they are passionate about.
One might think that after 16 cohorts and several hundred Breakable Toy projects we've seen it all, but nothing could be further from the truth. The surprises just keep coming, leaving us in a continuous state of awe. The creativity, passion, and technical expertise that the Launchers exhibit in their capstone projects is a testament to the tireless dedication they bring to changing their lives. It's about time we show the world what they've been building.
In the first of what will be a series of Breakable Toy showcases, we'll take a look at an app built by Filipe, a Launcher from our 12th Boston cohort.
In an era of polarized opinions and constant disagreement, hopefully there's at least one thing we can all agree on: language is a pretty important part of our lives. It'd be exceedingly difficult to express complex ideas and emotions without it. While breadth of vocabulary is not necessarily indicative of good communication, there's certainly a high degree of positive correlation between the two.
When it comes to music, there's obviously a considerable variance in the lyrical diversity of popular artists, but most of us would have a difficult time quantifying that difference. Wouldn't it be great if there was an app that could examine the lyrics of all of your favorite artists and show you a visual representation of the relative differences in the variety of words they use? Well, Filipe thought so, and he built an app to do just that. Dubbed Musicabulary, it's one of the most interesting uses of data visualization we've seen yet.
Quite an achievement to put it all together in just 2 weeks! Here's Filipe with the full overview:
(If the above video doesn't work, here's the direct link)
Musicabulary lets last.fm's users explore how their favorite artists compare to each other when ranked by the wordiness of their lyrics. It approximates an artist's lyrical diversity by counting the number of unique words within their first 3,500 words on record. This small sample allows it to make comparisons across artists, even when they have only released a handful of records.
If you have a last.fm account, you should sign in to take a look at your results. Unfortunately, if you don't, you won't be able to see your own personal chart quite yet. I'm working on that.
I don't aim to suggest that this metric is a good proxy for lyricism or any other artistic quality really. The results are heavily influenced by genre and decade, among a lot of other variables: hip hop artists unsurprisingly tend to come out on top, while musicians from the 50s and 60s tend to rank close to the bottom. When I look at my own results, i find that Dylan (943 words) ranks below the average and is almost 400 words behind The Menzingers (1,315 words). If there is anyone out there interested in arguing that The Menzingers' lyricism surpasses Dylan's, I'd love to hear from you.
Still, I find the results interesting as just another way of thinking about my favorite music. I figured others might as well. I built this project as a way to dip my toe into d3.Js. Playing around with last.fm's data was the icing on the cake.
This entire project is heavily indebted, both in substance and style, to the great work by Matt Daniels and other folks at Polygraph. I had the initial idea for the project after reading Daniels' post about The Largest Vocabulary in Hip Hop. I tried to follow his methodology wherever possible, but using 35,000 words as he did made load times Impossibly slow. It also excluded far too many more recent artists, especially those outside of hip hop. As a result of the smaller sample, I'm admittedly capturing a less representative picture of each artist's lyrics, leaning heavily towards their earlier discography. This is not entirely fair towards older artists, whose first couple of records may not be the records they are most remembered for. Nonetheless, I find that the results are still, to steal Daniels' words, "directionally interesting." Hopefully you agree.