Loud Rock at KGLT - playlist analysis and demographics

Ever since I began doing my radio show at KGLT, I've been collecting playlist data and coalescing it into a spreadsheet. I've been interested in analyzing that data to answer some questions about my listening and playing habits. Early on, I just didn't have enough data to evaluate, then life got busy and I didn't have time. Now that my steady show at KGLT is no longer, I wanted to analyze the data (over a year's worth) and see what it tells me.

I've often made the claim that a "lot" of the bands I listen to are from Finland and Sweden, however, I've never quantified that. That's one area I wanted to investigate using my KGLT playlist data. I also wanted to look at my data in the context of some of the "metal demographics" maps that are out on the web. They typically show Scandinavia, especially Finland and Sweden as contributing a disproportionate amount of metal to the world per capita (see Demographics of Metal). Of course, my analysis will only be one view of the problem, specific to the music I play at KGLT, but I still think it's interesting to analyze it in the context of the bigger picture.

My data is arranged to show every band I've ever played at KGLT and how many times I've played them. The simplest version of the spreadsheet sorts that data from most often played (#1 Nightwish, #2 Amaranthe) down to all the bands I've only played once. Right off the bat, I see hints of what I already suspected: in the top 10 most played bands, 3 are from Finland and 3 are from the United States. However, Sweden only has one band in the top 10. Looking at the top 20, Finland and Sweden each show up 4 times but the US is there 5 times. So, at the simplest level, yes, I do play a lot of Finnish and Swedish bands, but not at any obvious cost to the US.

However, the US starts showing up in heavy numbers in slots 21 and below, quickly establishing a higher play count overall than even Sweden and Finland combined. In fact, the US swamps everyone else with 462 plays, with Finland the next closest at 86 and Sweden following at 77. That, however, feels a bit distorted to me. I'd spent many months reviewing the entire KGLT Loud Rock CD collection and had played any band that got my attention. That collection is fairly weak in imports, so the pool I had to pull from was a bit skewed against Finland and Sweden. Many of the bands during that library review period I only played once. With that in mind, I decided to trim down the data set and regenerate my charts. I ended up with charts just like the original except restricted to 3 or more plays, 6 or more plays and 8 or more plays. Only the most restrictive (8 or more plays) chart levels the tables, showing Finland in the #1 spot followed by the US and then Sweden, but all in comparable numbers. Clearly, when push comes to shove, the music I like to play the most is, in fact, biased towards Finland and Sweden.

I also have claimed that I lean toward new music, playing lots of recent releases and bands from the 1990s and 2000s. To test those claims, I rearranged the data to look at number of plays per year the band was formed. I was pleased to see the 669 of the songs I played were from bands formed in 1990 or later. Only 223 of the plays are from bands formed prior to 1990. Being disappointed by the prevalence of "Classic Rock" stations and the dearth of new music on the radio, I was happy to see that I've been doing my part to reveal all the recently released heavy music out there in the world.

Having confirmed my suspicions at a very high level, I started to wonder how that relates to the population size that's generating these bands. I was highly suspicious that the main reason the US dominates the overall list is based on population size. I've seen the "metal demographics" maps and know how small the Swedish and Finnish populations are compared to the US. In addition, given my own knowledge of variation in metal style between states here in the US, and relative populations, I was curious about how things would change if I treated states as comparable to countries elsewhere in the world.

The first chart I generated considering states as countries showed something that won't shock anyone. The entity that generates the most plays on KGLT is California, immediately followed by Finland, Sweden and then the state of New York. Also, as no surprise to anyone, England and Germany land in the 5 and 6 slots. The next 10 are mostly states, including a number that I would have guessed (Massachusetts, Florida, Georgia and Texas). I don't know that I'd have guessed New Jersey and Illinois would be this high in the list, but Washington makes sense considering Seattle's contributions to music. Those of you outside the US will be pleased to see Canada at #8, the Netherlands at #12 and Australia at #16, immediately followed by Norway, Italy and Greece.

I was still a little surprised to see California dominating the list but once again remembered all those single-play bands from my library review efforts. We know that California, especially the southern part of the state, churns out a ton of bands, so it might still be simply the result of excess availability relative to imports. With that in mind, I decided to try looking at bands with higher numbers of plays. Even at 3 plays, the playing field levels quite a bit, with California leading at 88 but Finland in the same ballpark (73) and Sweden at 56. Being even more restrictive and looking at 5 plays or more pulls Finland into the lead at 63, with California at 61 and Sweden at 44.

I also wanted to look at the cities spawning most of these bands. Now, at this point my data is a little less robust, since the Wikipedia, where I gathered band information, is a bit spotty on how it describes where bands are from. Places like New York City and Los Angeles often are credited for bands that were formed in a borough or nearby city. Even so, I think the table is interesting, keeping in mind that Los Angeles and New York City probably have higher numbers than reality. It's definitely interesting to see Gothenburg as contributing the second largest number of plays and Helsinki in the #4 spot (note I'm skipping the true #1 in the list which is "undetermined").

With all those questions answered, I still had a nagging curiosity. How much do the large populations of California, Germany and England help them generate large numbers of playable bands (for my show). So, I made one more table, scaling the number of plays by the population of the state or country.It's impressive that California has so many plays that it still lands in the #4 spot despite its large population. Sure enough, Finland and Sweden, with their smaller populations and large number of plays land in the #2 and #3 spots, but what about #1? I had to laugh when I saw it. I love Tyr and a played them a number of times, but they had never generated an entry anywhere near the top of my tables before this. Thanks to the tiny population of the Faroe Islands, however, it lands in the #1 slot. Basically, I can conclude that the Faroe Islands, per capita, are the most effective country for generating music I like to play on the radio. Go Faroes and Go Tyr!!

In the end, the data simply reinforces what I already know. I love Finnish and Swedish metal. Those two countries are highly efficient at creating new bands that I like, and at the same time generate a wide variety of sub-styles within the metal umbrella. The United States and California, where I was born and raised, clearly generate a ton of listenable heavy music, and I should not overlook many parts of the US for their contributions to the overall metal pool. Sadly, Montana does not even appear in the data. Perhaps I can take small solace in knowing that Oregon does not appear in the data either, and Portland is currently a lively music center. So there's always a chance I just didn't get around to playing that one metal gem from Montana.

KGLT radio: biweekly routine of a late-night DJ

My life as a DJ started mostly on a whim. It just happened that I learned of KGLT's apprentice class about two weeks for the summer session started. The timing was perfect; there was no opportunity to think too hard before committing. Because of that, I really didn't have a motive or justification for becoming a DJ. It was a chance occurrence, an interesting opportunity that I couldn't pass up.

Looking back now, however, there are a few clear reasons that I continue to DJ at KGLT despite the hours of preparation and the late nights in the studio. To better express those benefits, I need to describe my biweekly routing as I prepare for and then deliver my show.

The cycle begins when I decide on a theme for the show. Sometimes that's easy, when my show is near Halloween or Christmas. Other times, I focus on favorite guitarists or bands that have influenced me. Some weeks I just decide to play whatever feels right. That said, in this article I want to describe an approach that has kept me going for months worth of shows. KGLT has their CDs and LPs filed in different categories, much like a public library. There's an impressive array of rock, jazz, blues, country, bluegrass, etc. There's also a small plastic bin for each style that houses the new releases. My category of choice is "Loud Rock" encompassing all sub genres of metal, plus some punk and heavier alternative bands. Although I have a good collection of music, KGLT's library includes a large number of bands and albums that I've never heard of. After the first few months when I was drawing mostly from my own collection, I reached a point where I needed to mix in a generous helping of new material. That's where my goal began...

I decided to make my way through every loud rock CD at KGLT and evaluate each one for use on my show. My goal was to accomplish that fairly quickly, making "Yay" or "Nay" decisions efficiently. That way I can evaluate enough albums in two weeks to prepare a show entirely out of KGLT material. The key step for me is spending 3-5 hours at KGLT previewing one CD after another. The first level of filtering is always easy. I'll listen to the beginning of a few songs, if any of them grab me, then I jump into the middle of a song, listen to some vocals and try to find a hook that gets my attention. The easy "Nay" decisions for me are songs that fail my "melodic" criterion. Generally that means musical forms with only guttural vocals (screams and growls) get nixed quickly, and that tends to remove about 50% of the albums I review with only a minute or two of review on each. Sometimes things bog down a bit, when I run across bands like In Flames, Amaranthe or the SLoT, where gutturals are used in combination with melodic vocals. I'm typically happy with a balanced approach of that sort, but quick scanning can sometimes give the wrong impression, since songs will often use gutturals at the beginning of a song and shift to melodies in the hook. Bands like that can often take more like 3-5 minutes. Melodic death metal can fool me in the opposite direction: the instrumental parts at the beginning of each song will sound great, sometimes reminiscent of heavier power metal bands, but when the guttural vocals kick in I realize the band is not in my wheelhouse.

Once a band or album makes it onto my "keeper" list, then it's a matter of picking a song. Sometimes I'll scan through the album in search of something that grabs me. Sometimes I'll look in iTunes for songs that have gotten a lot of listens or read through the band's Wikipedia entry to see if any songs played a key role in their success. Having picked one or a few potential songs, then I listen and pin down the song for my set. I keep track of band name, album, song title, track number and track duration in a form that I can quickly migrate into a spreadsheet.

After I've collected roughly enough songs for a full 3-hour show, I pull my data entries into a spreadsheet and start arranging my show. There are some restrictions that I need to factor into the playlist:

At the top of each hour, we have to announce the station ID and thank that hour's underwriters
Two other times during the hour we need to make public service announcements (PSAs) and play promotional clips
Factoring all that in, I've found that around 2 hours and 42 minutes worth of songs is what I need for a full show, roughly 54 minutes for each hour. Using that knowledge, I start arrange the songs in an order that works for my mood, trying to create three clearly defined blocks of around 54 minutes each. Within those blocks I try to find three groupings of 3-5 songs each, such that I have a PSA/promo slot around 20-25 minutes after the hour and another one at 35-40 minutes after the hour. I add some additional annotations to the spreadsheet to make the breaks stand out and make it easy for me to find the song, whether it's on my own (burned) CD or on one of KGLT's CDs. If I'm bringing any material from my iTunes collection, I burn those songs to 2 sets of identical discs. That allows me to segue between songs, even if I decide at show time to reorder some songs or skip others.

Once I've prepared my playlist spreadsheet, burned any discs from my personal collection and collected any of my own CDs that I have readily available, then I set everything aside until its time for my show. Typically I reach the KGLT studio around 11:30pm. Most nights, Adam (Kish) is in the studio when I get there. Before I do anything else, I go through the new Loud Rock bin and the Loud Rock drawers, pulling any CDs I need for my show. I stack the CDs in the order I plan to play them, inserting my own CDs as needed. After that I chill for awhile and catch up with Adam. We often talk about metal, Iron Maiden, Thin Lizzy, etc. It's always fun talking to someone else that loves heavy music, has eclectic taste and likes talking about it.

Then, somewhere around 11:55pm, Adam starts playing his last song of the night. He clears out his gear and CDs and he submits his playlist (more on that shortly). Then I move my stuff into place, near the 3 CD players and the console. I fill out some paperwork needed from all DJs as part of having a show at KGLT and put the first two CDs I plan to play in the empty players (Adam's last song is typically in one of the players). I get the songs queued up, so that I'm ready to go. Usually Adam's last song ends right around the top of the hour and I'm on my way.

In the last minute before my show, I switch to headphones and activate the mic but leave it potted down (i.e. the gain is all the way off). I also bring up the level on the channel for my first song, usually around -5 dB on the board. Then, just as Adam's song is dying away I bring up the mic gain, read off the station ID, announce the weekend weather and thank the underwriters. The weather information is available on the main screen of the in-studio computer (a Windows PC); the underwriter info is displayed on the side screen of the same computer. At the end of my announcements I hit the play button on the console and start the music.

The console makes it easy to listen to the beginning of each song prior to playing it, using the Cue channel. So, as one song is playing, I'm listening to the next one and deciding whether a slow crossover fade is in order or the next song needs to kick off on a dime. I segue my way through the first block of 3-5 songs, while taking care of another KGLT DJ responsibility: entering band names, album names, song titles and "new or not?" into the playlist app on the studio PC. We need to track all the songs we play so that the radio station pays appropriate royalties. Those same playlists are also posted on KGLT's website. Then, as I'm playing the last song before a PSA break, I once again switch to headphones, get the mic ready and wait for the song to tail off. Then I read the PSAs, announce the last block's songs and fire off the next song.
The pattern repeats for my next block of songs, although perhaps this time I'll use prerecorded PSAs and promos, which are accessed on the studio PC. I still announce the prior block's songs myself and then kick of the last block of the hour, which takes me back to the top of the hour. At that point, my routine is very similar to the beginning of my show (station ID, announce underwriters) but I don't do the weather, and I do announce the songs from the previous block. That second hour and the third hour are very similar except as my show winds down after the third hour.

At the end of my show, no one else is coming in until 6am. As a result, less than a minute before my last song ends, I cue up SAM, the computerized DJ and crossfade at an opportune moment. SAM automatically cycles through songs, plays PSAs and promos, and carries over until the morning DJ comes in at 6am. SAM uses voice synthesis, and she sounds a little strange, but she does a good job filling time and she's happy to tell you that she plays "some smooth jazz". Once SAM is running, I refile all of the KGLT CDs I used. I also burn copies of my show to CD to take home (that way Nancy can hear my show without staying up). Then it's time to turn off all non-essential lights, back down the studio volume and the two other speaker systems in the facility, pack my bags and go home.

It's often tough to get to sleep after my show. I'm usually pretty amped. Often, as I'm trying to settle I start having ideas for my next show. And thus, the cycle repeats itself.