The Capital of Metal?

Back in 2018 a campaign was set up to “determine which Finnish city deserves to be known internationally as the Capital of Metal”, stemming from Finland’s fame of being the “most metal” country in the world based on bands per capita. They did this by spreading the word and getting people to submit bands to be considered, with prizes and clout for the bands and city that would win. The small municipality of Lemi ended up winning the title by a fair margin.

For me, this fun little exercise illustrates some issues and questions that constantly bounce around in my head when looking at analyses of metal vs population. Below I have incoherently vomited outlined various thoughts on their campaign and how it compares to information available on Metal Archives.

Let me just pause here and say of course this campaign was for fun/publicity/etc. and my considerations and questions are also just for fun. So, don’t get me wrong, I am probably overanalyzing here, but I find it very interesting. Wh…what? Stop looking at me like that!

Intended Goals and Qualifying Metrics

To begin with, let’s look at the stated goal of the campaign and the selected metric for achieving that goal, namely: what is the “most metal” city/municipality in Finland as determined by number of bands per capita.

From the start, what do we mean in terms of a city being the “most metal”? I think often times what this actually translates to in our heads is either: where is the highest density of people that are into metal or what is the relative cultural/economic output of a locale compared to is population. Band counts give a reasonable approximation to these questions, but I would think musician counts, album counts, records sales, etc. would better answer these questions.

Another question is one of time. Are we asking where the most metal place is today, of all time, on average, in recent times, etc. and how do we create our metric from raw data? For instance, do you count all bands of all time compared to today’s population? Bands formed per year by yearly population, then take an average over time? Should we only include more recent “steady state” years and not include the initial growth years?

I really don’t think taking all bands over all time and comparing to today’s pop makes much sense, but it tends to give similar results to averaged yearly numbers.

What is a “Band”?

What really determines what counts as a “band”? There are a number of problems here.

• Bands can have a varying number of people and can often share members.
• Bands can form and disband without actually producing any music or as little as singles or demos. It often takes a few years before a band releases any material.
• Should we count bands that are not or have not been active for decades, especially when comparing to today’s population?

Example issue: Broken Mirror

• Broken Mirror from CoM doesn’t seem to have released anything.
• There is a different Broken Mirror on Rate Your Music
• There is yet another Broken Mirror on MA

Ultimately, I think musician counts or album counts are better metrics as they tend to be monolithic values.

Data Collection and Classification

The campaign’s method for gathering band counts is fairly reasonable. A set period of time is given for people to submit metal bands. Compare this to Metal Archives, which has been gathering information for decades, and you will get a smaller number of more popular or recent bands vs. a larger number of more obscure bands respectively.

There is also the question of what is a “metal” band? Well, that’s an entirely subjective question and you will see that many of the bands counted for CoM may not be included at MA. The results however may likely be similar as there tends to be at least a reasonable consensus on what makes a band metal. Ultimately, it all comes down to subjectivity and neither answer could be considered objectively definitive.

Example issue: counts for Lemi at CoM vs. MA. CoM has 13 bands vs. MA which only has 5! This is a significant discrepancy especially with such small numbers which can change the ranking significantly. Which one is more “correct”? Kinda both, and a reason to view either result as non-definitive.

Small Numbers

One factor working in Lemi’s favor is it’s relatively small population. This can cause issues as a result of the resolution of the bands per capita metrics. Each single band in Lemi is worth ~33.76 bands/100k vs. in Helsinki where each one is only 0.15 bands/100k! This means that a high total bands/capita value in Lemi is more prone to be noise rather than signal of “metalness”. In other words, over a long period of time, you are more likely to maintain an average bands/capita/year in Helsinki than Lemi.

Another good example of this is the municipality of Hailuoto. This small city of less than 1000 residents has a whooping 105 bands/capita! This is a result of just one band. If we were to look at this over time, this would only look like a blip in time as opposed to a trend. Averaging over time, bands/capita for this municipality would likely trend closer to 0. So how would we use this info for a ranking?

Metal Archives Data

Here is my attempt at using MA to find the “most metal” municipality in Finland.

Notes:

• Includes all bands over all time at MA
• MA uses old town/municipalities so I had to translate some of the numbers.
• There are 500+ bands with either no or insufficient location information.
• If a band is from multiple locations or changed locations it is counted for each location, there is some double counting.

Other Notes

Something a bit tricky with the data is how Finland distinguishes towns, cities, and municipalities. CoM seems to call what it is counting “cities”, but from what I can tell these are municipalities according to Wikipedia. Wikipedia also tells me that these municipalities can be identified as either a “city” or a “town” but may at the same time be amalgamations of multiple villages or suburbs. I imagine this distinction as being akin to counties in other countries. Regardless, we’re looking at designations from Wikipedia that roughly match what CoM has.

Lastly, I just want to say I am not Finnish, a “real data journalist”, or a “real statistician”. I’m just doing this as a hobby. So, if any of this seems asinine, poorly worded, etc. there’s your answer why.

Still reading? Follow me on Twitter!