Last year, I made a series of visualizations based on playlist data from my favorite local radio station: KEXP. Well, it’s a new year and there’s new data, so here I go again! First, let’s start with the classic viz I made last year. This view shows the top artists, albums, and tracks of the year. The color denotes which DJ played them. You can use the dropdowns to view specific DJs or toggle between artist, album, and track.
Our big winner here was Spoon, not surprising as they released one of the best new albums of 2014. However, DJs must’ve been drawing from their back catalog a bit to get them up to all those plays because their new album was in 5th place when it comes to top albums. The top spot actually went to what was probably my favorite album of the year, “Sun Structures” by Temples.
I mean…that hair + that awesome psychedelic sound….what’s not to love?
By the way, if you look at the top albums and tracks, you’ll often see an asterisk for artist. This is because the DJ assistants at KEXP who input all the playlist data can’t agree if it’s Alt-J or alt-J. Capitalization irregularities actually plagued this whole dataset, so some of the numbers may look a little lower than they should be. Another issue is when bands have song titles with the same name. For example, a track named “Feel” appears to be the second most played song, when in actuality that’s a combination of plays for songs by Bombay Bicycle Club, Ty Segall, Big Star, and a few others. Unfortunately, you can’t use combined fields in parameters in Tableau, so there was no way for me to fix it on this particular dashboard.
The number one song is “Red Eyes” by The War on Drugs. I was especially tickled by the high placement of the song “Queen” by local hero Perfume Genius. This song is amazing and deserves the number spot for the line “No family is safe when I sashay” alone. The video is pretty weird and rad:
The whole album is amazing and I couldn’t help but use intoxicants and lay on the floor and listen to it when I put the vinyl on for the first time.
I wanted a couple dashboards that people could go into and make insights about their favorite bands and DJs. First, an artist explorer. You can use the dropdown to choose a couple different artists to compare. There’s a text search, if you don’t want to read through the whole big list. As a starting point, I compare my two favorite Seattle bands, La Luz and TacocaT. They seem to be pretty even in plays.
Next, here’s a dashboard that looks at what DJs like to play. I’ve started out with infamous morning DJ and tastemaker John Richards. Man, that dude loves Strands of Oaks’ newest album. Especially on Tuesday mornings at 6 am. In fact, he actually said himself it was his favorite album of the year. It’s pretty fun to click on one of the artists on the right and see what time they are played the most.
This year, I really wanted to do something with the top 90.3 albums of the year. KEXP had a form up on their website in December for their listeners to vote for their five favorite 2014 album releases of the year. They tallied up the votes and ranked the albums and did a fun countdown at the end of December. I wanted to compare the ranks by listeners to how often those albums were actually played. That’s how I got this nifty quadrant chart:
I’ve divided the chart into 4 sections. “High Listener Rank, High DJ Plays” means that the listeners and DJs were in agreement that these albums fucking rocked and should be considered to top honors. The listeners and DJs are also in agreement in the “Low Listener Rank, Low DJ Plays” quadrant. The interesting stuff is really in the other two quadrants. I noticed that in “Low Listener Rank, High DJ Plays” there are a few more local bands and just generally less known bands. This section is showing our taste-makers at work. I’d bet that a lot of people voted for these albums after hearing them first on KEXP. The “High Listener Rank, Low DJ Plays” section is interesting because it’s filled with indie favorites: The Afghan Whigs, Jenny Lewis, Aphex Twin… these are some heavy-hitters. These are all pretty well-known artists outside the college rock scene, which is maybe why KEXP plays them a little less than some other bands.
Last up, I just wanted to make some quick points of things I thought would be interesting to know. For example, did you know that the artist with the most distinct albums played on KEXP was Johnny Cash? It probably helps that he made a buttload of them. Click through these story points if you want to see more stuff like that.
I think the third story point is especially interesting. I filtered the list to only show artists that only have one song played on KEXP and than sorted it by number of plays. So you can think of it as the ultimate 1-hit wonders on KEXP in 2014.The song that ended up in the top spot on that list was actually pretty cool:
Dig into those dashboards if you feel so inclined. And be sure to tweet me and interesting tidbits you find in there!
I’ve been trying to come up with a pun all morning, something to do with Beat Connection and Data Connection, but it’s too early and coffee hasn’t happened yet.
In honor of our big company party featuring a performance by my favorite local band, Beat Connection, I’ve made this dashboard on how much I’ve scrobbled plays on Last.fm. Of course, it’s a little less than how much I actually listen…. It doesn’t count all the vinyl plays I do:
KEXP is a local public radio station here in Seattle. It’s an awesome station and they play lots of indie and lesser known bands. They stream online, so everyone can enjoy their awesomeness. They also keep a live-updated playlist that shows every song they ever played since 2001. Pretty incredible and kind of a data-slash-music-nerd’s dream (and definitely worth supporting!). I used import.io to scrape everything from 2013. I’m going to work on getting the other years in there, too.
I had all kinds of grand ideas on what kinds of things I would do with this data. I made a lot of crazy looking charts. But when it came down to it, the simplest charts told the best stories. So I scaled back. What ended up being what I think is the most interesting dashboard with this data is pretty plain. It’s simply how many plays a dj played an artist/album/song.
I was playing with some of the time-based data as well. When I did a plot of Week vs Song Plays, it’s pretty obvious to see when KEXP is doing a pledge drive; number of songs played drops over 15% each time. I also looked at the weekly schedule by making a heat map of days of the week versus hours in the day. You can see blocks where they play less music: like Saturday and Sunday mornings when a local news talk show plays instead. The heat map under that shows the number one artist by plays for each hour of each day. Apparently 12am on Thursday is a really good time to hear some David Bowie.
I’ve been wanting to make a Phish dashboard for a while. I wanted to use the phish.net API and collect all sorts of delicious, raw data of setlists and locations and create all kinds of things. But alas, it doesn’t look like I will be allowed to. I gave up for a while, but then this tweet popped up in my feed:
I took the data, which looks identical to phish.net’s song list, and put my spin on it. Not quite what I was hoping for, but I think it’s kinda neat. I do appreciate that the “Cyclic” color theme almost matches the Phish logo perfectly.
Sasquatch! Music Festival has a killer lineup this year. It’s no surprise that tickets sold out in a record 90 minutes. With tickets going that quickly, people are turning to the after market to try to score wristbands. All over the forums, ticketless fans were repeatedly told to wait until a couple of weeks before the festival, that way ticket prices will be close to face value. These graphs were floated around a lot:
The pattern this year, however, is a little absurd. I’ve been scraping the Seattle Craigslist for posts and I put the data in a Tableau viz. As you can see, ticket prices have not had any kind of consistent downward trend. They are rarely offered for below $400, even now that we are only weeks away from the Memorial Day Weekend festival. In my personal experience, half of these posts seem like scammers, too. One in particular keeps saying that they will only deal over eBay and asks for a bunch of personal information so they can “set up the transaction.” I bought my tickets presale, so I’m all set to go (thank god!), but I’ve been looking for tickets for my friends and it has been pretty rough. I’m not sure if it’s just increased demand that is leading to ridiculous after-market prices on Sasquatch tickets or if there are other factors, but I think we can all agree that these kinds of scalpers totally suck.
UPDATE 5/21/2013: I’ve been updating this viz basically everyday which involves cleaning up the dataset. There’s only 3 more days until Sasquatch and judging by the amount of “WANTED!” posts I had to delete from my data, I’d say it’s definitely a Seller’s market right now. Perhaps the age old wisdom of “Wait until the last week and prices will go under face value!” is proving to be obsolete.
I was once again exploring the last.fm playground when I viewed the following graph. The graph charts my top 90 artists of the last 12 months based on the average age and gender of its listeners. Apparently, I listen to music that is very much in my age group. Something I found interesting was the gender split. It really seems to correlate with the two genres I primarily listen to: more folksy things tend to have more female listeners and more electronic things seem to have more male listeners.
Back when I was making my Grammy’s dashboard, I was spending a lot of time on Last.fm. They have a lot of cool visualizations in their “Playground” section. One of them includes this slick looking area graph based on the mood of the music you’ve scrobbled:
I showed it to a coworker and he pointed out the pattern of listening to sadder, slower music every couple of weeks. He asked me what that was about. I couldn’t answer him right away. When I got home that night, I stared at it a little bit more. “No…this couldn’t correspond with….” I thought to myself. I pulled out my phone and opened up the app I use to track my menstrual cycle (yes, they make apps for that and if you have a uterus and are of childbearing age, you should use one) to see if the dates corresponded. Indeed, November 19th and December 22nd were both the start dates of my menstrual cycle. Weird.
This isn’t particularly useful information for me, but it’s interesting. You learn all kinds of interesting facts about yourself when do self data collection. My partner suggested that it could be used to predict when shark week is coming. It does seem that mood starts changing a few days prior. Now he knows what to expect when I start listening to John Coltrane in the bathtub.
This is the first in a new series I’m doing called Pop Viz, where I’ll explore data related to pop culture. This first visualization is the first one I made after starting my new position at Tableau. I looked up some last.fm data for Grammy nominees.
Some of my favorite highlights from the dashboard:
Best Dance Single nominee Al Walser appeared out of nowhere, receiving the nomination after heavily courting the voters. If you look at Al Walser’s weekly listener trends, he shoots way up after the nominations were announced in December, presumably because no one had heard of him before then.
Way more people listen to Michael Buble in the Christmas time. Coincidentally, he’s nominated for his album entitled “Christmas”
I calculated the amount of scrobbles per listener for each artist. Artists with lower scrobbles per listener tend to be nominated for song categories, while those with more scrobbles per listener are more likely to be nominated for albums.