The Rise And Fall Of Prog, Part 2

In my last essay I discussed some evidence in the Progulus Radio data that shows a general decline in the number of highly rated albums over the past few years. As promised I did take a closer look at the data.

Let me begin by stating that others have made a valid point that my statistics only reflect the popularity of certain artists and albums by the listeners of Progulus Radio who are a distinct group of proggers and may not reflect the tastes of others within the prog genre. Prog music by it’s nature was never meant to be popularized like traditional music. We are all creatures of habit, and what's popular on one prog radio station might be vastly different on another, or people' personal listening habits, or other prog groups or forums groups. I personally enjoy listening to many of the obscure and offbeat albums and the extremely wide variety of music that this genre offers. Musical tastes are a very personal thing, and the true progger should look beyond what is the flavor of the day and use the statistics only as a guide relative to their own personal tastes.

When I took a closer look at the ratings and play data, I did find a general decline in many stats from about 2005 onward. This includes the total number of albums released each year, the number of highly rated albums, and the amount of song play for albums released on a certain year. The problem is that I'm not really sure what it means. Since this is the first time I've studied the data at this level I feel like I've only taken a snapshot of a dynamic system. The song rating and airplay stats are constantly evolving and changing over time, so it's difficult to say with certainty how things will look a few years down the road. I know that personally I often find myself re-rating songs differently than when I first heard them.

My study did find one certainty though. The number of albums per release year on Progulus has steadily been on the decline since 2006 as the next graph will illustrate:




This seems to indicate that the industry has considerably downsized recently and may continue to contract. We may find in hindsight that 2005-2006 were the peak years for the third resurgence of prog. Additionally, by correlating data between the number of releases per year and the number of highly rated releases, I found a connection that suggests that the more releases that are put out the more highly rated ones we’ll have. But sadly since the proliferation of new albums has been steadily decreasing since 2006, I think we need to accept the fact that we're not going to have as many great CDs per year as we have been graced with in the past. On the upside though I think the state of the industry is still a good one. There are many good artists who contiue to realease high quality albums so we should be all thankful for that. Interestingly the average song rating has increased slightly since 2006.

As a footnote I found the last graph pretty interesting and wanted to share it. This graph is shows the number of song plays per artist ranked from #1 downward:



This graph from left to right shows artist #1 with the most plays, followed by #2, then #3, etc. While it’s not really related to my discussion of the decline of the genre, it does show the loyalty of the listeners to a select group of artists. I wanted to integrate the graph but I guess there's not an easy way to do this that in Excel. It does look like about 90% of all the airplay on Progulus is from 500 artists of the 2,000+ total on the radio station. That is to say about 90% of all the artists could be eliminated with very little overall impact on our typical radio play.

Statisticians out there may find the following information handy:

Artist Airplay (number of song plays by artist):
Mean 121.6476075
Standard Error 7.551491308
Median 32
Mode 1
Standard Deviation 374.998269
Sample Variance 140623.7017
Kurtosis 594.2007154
Skewness 19.18907909
Range 13069
Minimum 0
Maximum 13069
Sum 299983
Count 2466

Album Airplay (number of song plays by album):
Mean 78.22503259
Standard Error 3.219954059
Median 43
Mode 11
Standard Deviation 199.4033084
Sample Variance 39761.67939
Kurtosis 2261.137573
Skewness 42.19692306
Range 10893
Minimum 0
Maximum 10893
Sum 299993
Count 3835

Update 14-Dec-09:

Since it has been asked for, I've included the following graph which correlates the number of good albums relative to the number of total albums each year: