Measuring song impact - h-FAWM-index

Skip to the bottom

  • @lowhum  Mar 7

    I read a lot of discussions here about quantity vs. quality, etc. and since some of my work (librarian) involves scientometrics I would like to adapt a metric used in the scientific world to measure FAWM impact (hopefully both Qs:)).
    It could be called the h-FAWM-index - deriving from the h-index and for those with a science bend - you can read about it if you like - I'll explain it quickly on an example.

    In the scientific world instead of songs you have publications (usually journal articles/books) and instead of comments you have citations (people mentioning/commenting your works).

    So I'll jump directly to the metric - if you want to measure your h-FAWM-index - take all the songs and line them up in a descending comments order. The h-FAWM-index will be the number of songs that received at least h comments.

    E.g. you have 4 songs and you have this situation:

    song 1 - 40 comments
    song 2 - 26
    song 3 - 5
    song 4 - 1

    So your h-FAWM-index is 3 (i.e. you have 3 songs with at least 3 comments)

    It's just a curious thought but might give some food for some!

    If you struggle with the calculations - send me a note and I'll try to calc it for you!

    PS: It is still debatable in the scientific world that anything can measure quality, safe along in the artistic realm of creation, so take this lightly as an Apollo's stringer on Pan's flute πŸ˜€)
    PS2: I forgot to mention that this index tends to grow, so don't despair if you're small today πŸ˜€)

  • @adforperu  Mar 7

    So am I right here... I've got 15 songs, all over 15 comments - so index = 15?

  • @devin  Mar 7

    My dataset:
    39
    35
    26
    22
    17
    16
    16
    14
    14
    13
    12
    11
    10
    7
    7
    7

    So I have 11 songs with 11+ comments, yes?

    Would it take more pondering to discover what the true correlation is?

    h-FAWM-index seems like an interesting way to weigh the comments per song, however, comments are a function of more than song quality.

    How many comments I give will affect how many I get, as well as how many long-time friends I interact with here, plus any promotion I do, and existing fans, the collabs I do, and if I behave (or misbehave, heh) in the forums.

    Maybe?

  • @oddbod  Mar 7

    22 songs, the least comments is 10, so......
    ....er no,sorry I'm confused

  • @johnstaples  Mar 7

    @lowhum you said "So your h-FAWM-index is 3 (i.e. you have 3 songs with at least 3 comments)"

    did you mean you have 3 songs with at least 4 comments? where h=4?

  • @lowhum  Mar 7

    @adforperu - right πŸ˜€
    @devin yes = 11, - yeah - comments are much more than quality - i just tried to keep it scientific for fun πŸ˜€) Quality is hard to measure, ain't it? Who read the Zen and the motorcycle?
    @oddbod - you've got so many comments, man - you should be celebrating, not calculating πŸ˜€) -anyway - your index = 17 (17 songs with at least 17 comments)

  • @lowhum  Mar 7

    @johnstaples h = 4 means: 4 songs with at least 4 comments. Hope that clarifies - if you're wondering - send me your data πŸ˜€

  • @tcelliott  Mar 7

    @lowhum I reached H-Fawm-index of 14. Would you confirm (28 songs total):
    29
    28
    23
    23
    22
    22
    18
    16
    16
    16
    16
    15
    15
    14
    14
    14
    14
    13
    13
    12
    12
    11
    11
    10
    9
    9
    7
    6

  • @johnstaples  Mar 7

    @lowhum I am confused. Do you take your total song count and use that as the value for h? and then is is how many songs are at least h or at least h+1?

    I have 20 songs and my comments are,

    68
    38
    33
    32
    31
    30
    26
    23
    22
    22
    20
    20
    20
    19
    17
    12
    11
    11
    10
    8

  • @lowhum  Mar 7

    @tcelliott you have perfect control over the metric πŸ˜€) I just hope 14 means something to you! πŸ˜€) Greetings!

  • @oddbod  Mar 7

    @lowhum I’ll take your word for it but I still have no idea how you get to 17 πŸ€”

  • @johnstaples  Mar 7

    I think I get it now. So mine would be 15?

    https://en.wikipedia.org/wiki/H-index

    Start at top and count as you go down the list. When your count is greater than the number of comments on that item back up one/subtract one from your count!

  • @johnstaples  Mar 7

    @lowhum what does this mean in extreme situations like this,

    song 1 - 5
    song 2 - 5
    song 3 - 5
    song 4 - 1
    index = 3

    song 1 - 5,000,000
    song 2 - 5,000,000
    song 3 - 5,000,000
    song 4 - 1
    index = 3

    both are the same?

  • @lowhum  Mar 7

    @johnstaples - yep - you're 4th one was a flop - and you know it πŸ˜€)

  • @johnstaples  Mar 7

    This is very interesting and fun! But we cannot take it too seriously since comments here are not subject to any sort of qualifications like a peer-reviewing and the time constraints and the effect of your commenting, etc. Still fun.

  • @johnstaples  Mar 7

    @lowhum yes, in the second example it is glaring but in the first one not so much, right?

  • @lowhum  Mar 7

    @johnstaples It is, but its pretty unlikely to find such extremes - then there is a g-index that takes care of some of that... but that's already a song in there, right?

  • @metalfoot  Mar 7

    My h-value is 16, I suppose. πŸ˜€

  • @tcelliott  Mar 7

    By my calculations, @metalfoot, you are correct.

  • @lowhum  Mar 7

    @metalfoot In the scientific world that usually gets you an Assoc. Prof. degree πŸ˜€) A couple of PhDs and a lot of government funded research! I wish we could find a government to finance FAWMs and similar endeavours!

  • @metalfoot  Mar 8

    @lowhum That was certainly one of the career paths I had open to me... but I'm happy with not having to publish-or-perish. 😁

  • @scottlake  Mar 8

    Interesting idea but you might find this more as a skirmish participation index than some measure of impact

  • @andygetch  Mar 8

    number 9, number 9, number 9.......

  • @lowhum  Mar 8

    @scottlake - you could probably exclude the skirmish ones - or add a skirmish normalization coefficient (SNC = 0.36?)

  • @sph  Mar 8

    Interesting topic. In the end it's just another way of exploring.
    Some time ago there was an interesting article about the connectivity of FAWM. Here it is:
    https://slackprop.wordpress.com/2017/...
    @andygetch 😁

  • @bithprod Mar 8

    My main question is, like always, Why? This serves no useful purpose at all. Let's say that you want an objective number for the quality of your songs (/lyrics/instrumental). This is what you need to take into account in addition to the comment count:

    1) The number of songs you've written (fewer usually means more comments per song).
    2) How many comments you've made (β€œThe love you take is equal to the love you make”), and how many different people you've left comments for.
    3) When each song was posted (late in the month = fewer comments).
    4) The musical/lyrical ability and sincerity of the person leaving the comment ("Dear son, you are better than Dylan!").
    5) How visible you are overall, like on social media, profile pages and this forum, as well as your credentials.
    6) Whether you've posted a full song or "just" lyrics or an instrumental piece.
    7) Your production, as a full-blown radio-ready production will draw more attention, and rightfully so, than a Band-in-a-Box demo.

  • @bithprod Mar 8

    8) Whether or not a comment is positive, negative or neutral in nature (which is subjective, "Sh*t's weak, dawg!").
    9) Your intended audience, and your genre/style. Free jazz is less likely to get a lot of comments than singer-songwriter stuff on FAWM.
    10) Related to 9), your tags. They are part of how you market your song, and some people are better at using all the right tags. Also marketing in general, like mentioning your song on the forums when you finish a challenge, skirmish, etc.
    11) Whether or not you are a performer, as your number of "fans" will be somewhat proportional to your number of comments. Also, extroverts are generally better at marketing themselves.

    And so on and so forth.

    I suppose it could be fun to make an algorithm based on all these factors, but in the end all that matters is whether or not your intended audience like your song. It's all a matter of reaching those 0.1% and nourishing that relationship.

  • @coolparadiso  Mar 8

    I think its an interesting metric - but doesn't really mean anything useful here for me because of to many other variables . You would have to group your songs in subsets that had similar scenarios to be useful as a selection procedure. A good example is that my highest commented song was a trite fun number about Fawm done in a skirmish - it would have no relevance out of Fawm. Wheres as a new Fawmer i new no one when i did my first few songs and got few comments until i started providing a lot of comments to others and also started skirmishing. It could possibly have more relevance for a return offender who has a more level playing field.

  • @yam655 Mar 8

    @bithprod The why was covered in the original post: "It's just a curious thought but might give some food for some!" It's the same reason as "why write X songs in Y days?" Ultimately: curiosity.

    My total audience approaches one for even my best work. It is not equal to one, but merely approaches one. My work this FAWM (and, indeed, this whole year) was designed to be fed in to a computer system to generate songs for me, in my style, based upon a large corpus of my existing work. (Well, not _completely_ generate the songs, so much as to be my co-writer. More. Faster. Better.)

    @scottlake You bring up an interesting point: Impact versus skirmish participation. Skirmish participation can be seen as increasing the impact of a person's songs, as plenty of newer FAWMers gain fans and watchers through the skirmish process. You can't simply discount skirmish songs, as some folks _only_ participate in skirmishes. I really think papers could be written about the intersection here.

  • @tcelliott  Mar 8

    At best it's an indication of half of your fawm reach or connectivity. But you can see that with the total number of comments given to received (not the ratio) and even then it can vary on the number of songs. I posted 28 and definitely had a higher total number of comments than if I had posted 14.

    If taken as fun, I see no harm as long as it's understood that it really doesn't mean anything in whole.

  • @scottlake  Mar 8

    @yam655 you are right about the skirmish thing and participation or not. Personally I’ve never participated in one and been doing FAWM for 11 or 12 years. It seems the quantity of them has sharply risen over the past few years to where the latest forum post thread got overrun by them during the last few days of February. The index could actually explode if a FAWMer writes just one song, as part of a highly attended skirmish with let’s say 30 participants. Wouldn’t they get an index of 29 or so?

  • @yam655 Mar 9

    @scottlake They would need to participate in 30 skirmishes, and have each of those songs get 30 comments to have an index of 30.

    If you mostly kept to yourself, but participated in one super popular skirmish and so have one song with 100 comments, you could realistically have an index of 3.

    If you participate in a lot of skirmishes, you're an active part of the community and there's a lot of mutual commenting, so... You would naturally have a higher impact than someone who mostly keeps to themselves, right?

  • @darcistrutt  Mar 9

    I had nine songs with nine or more comments so, if I’m understanding, my H is nine. Nine is a nice number. Though if you believe the children’s song seven ate (8) nine so it is no more...and seven is very full. 😜

  • @zecoop  Mar 9

    I just want to let it be known that I come to FAWM to get AWAY from math... I'm so offended. πŸ˜‰

  • @rainchaser  Mar 9

    Thought we always use math when we set odd time signatures, and polyrhythms. As well as counting the amount of beats in a bar. Removing the amount of beats in a bar. Adding the amount of beats in a bar. Adding count ins to start a song for recording. Adding the amount of sections to fill one song. Removing the amount of repeats/sections in a song. As well the type of notes to use in a bar, whole notes, half notes, eighth notes, quarter notes, sixteenth notes. Taking an aspirin after going through all of this. πŸ˜‰

  • @tcelliott  Mar 9

    Wait. There's math in music? Dang. I guess I have to find a new hobby.

  • @andygetch  Mar 9

    two FAWM's in a row at number 9, last 50/90 eeked to 10, there is something curious here........

  • @owl  Mar 10

    not 100% sure i understand this, but I think I have an index of 9 but if i were to delete my least "popular" track, it would jump up to 15. One outlier can really drag you down!

  • @yam655 Mar 10

    @owl, having unpopular tracks doesn't particularly hurt, provided that you have enough popular ones. You start off by sorting them by the total number of comments, so your unpopular 16th song doesn't change your index of 15.

  • @cmharwood89  Mar 10

    As somebody whose career is made or broken by h-index scores, I am tickled by this. I like it much more than the "real-world" h-index.

  • @elliottlawrence Mar 10

    This is neat. So if you have 8 songs with 7 comments and 6 songs with 8 comments, does that mean your h-index would be 7? (at least 7 songs with 7 comments)

  • @ianuarius  Mar 10

    It would be interesting to have a group of people listen and read through every song in fawm and then see which songs they commented on.

  • @quork  Mar 10

    @lowhum , I’m sure it wouldn’t be too much trouble for you to analyze every FAWMer and determine the average h-score. With a 9 rating I look forward to seeing where I fit on the scale.

    Thanks!

  • @kahlo2013  Mar 10

    My FAWM h-index is 13
    My work biomedical publication h-index is 30
    I guess I should stick with my day job!

  • @tcelliott  Mar 10

    @owl - As @yam655 mentioned, you have a H-Fawm-index of 15. It took me a minute to see what process was used to get the number.

    If anyone struggled to see how to get the number, this is my process.
    I copied my song list to excel deleted all the unwanted information (such as song name etc.) until I was only left with the comment numbers. I sorted from largest to smallest to organize those comment numbers. Then I scan down the list until the comment number is lower than the song number I go back one song number and that's the index #. So far it's worked with all of the examples I've tried.

  • @owl  Mar 10

    @tcelliott @yam655 ah I think I missed the "sorting by largest" instructions. Thanks. I'll put it on my CV when trying book gigs, I'm sure everyone will be very impressed πŸ˜‰

  • @lowhum  Mar 11

    @owl Yeah, I guess for gigs the metrics should be how many people showed up... πŸ˜€) Then we'll have Beatles and Stones competing with the likes of Einstein and Freud πŸ˜€))
    @quork Yeah, I guess soon it will be "the data could and will be used against you... "
    @kahlo2013 30! Congrats! - You must be at least an Assoc. Prof. at some fine institution or other πŸ˜€) I'm already humming an h-index song with an enormous impact factor!

Leave a Message. Log in to FAWM or sign up first...