What do we want to see in the Playground?

  • Popularity graphs of artists over time.

    • Indeciso sa...
    • Användare
    • 25 jan 2010, 05:05
    I love the idea of yearly charts. A breakdown of each year on one page. 2009's top artists, top tracks played, top albums, what new artists you listened to the most, etc. All these stats would be made even better if you could see them side by side to past years. Now that I've been here since 2006 I've been wanting this so bad lately. If this breakdown could be added to months also that would be great but yearly charts are better to me.

  • all-time (not weekly) loved tracks

  • I'd like to see a cleaner layout, more profile customization (Choose profile backround colors?), and a way to change your username.

    Also, I would love to have a way to have your own personal links/bookmarks to areas of the site. For example, having a bookmark to an artist's page, or a community page, or a site feature; would be really cool and help us to navigate the site quicker

  • Random radio.

    This radio would work like this:

    The user would click on play and the radio would start to play a random song from all songs on last.fm database.

    After the song finish, the user would give a rate to the song the song from 1 to 10, then the radio would continue to play songs in a totally random way play and he would rate those songs played.

    After some time, the last.fm, based on the rates of the users that used that radio, will be able to start to give recommendations to the users and the radio will start to play songs based on that recommendation.

    If there is still no data to make recommendations to this is user, the user will still continue to get recommendations.

    PS: This radio would have a minimum amount of random songs, so if the first songs that the user get were already rated by many users (and the system can give recommendation based on those few songs), he will still need to listen to more songs.

    There is tons and tons of songs on last.fm, it will take a extreme amount of time to create a database of rates needed to make recommendations in this random way. Don't you think that the users will need to wait a extreme amount of time before they start to get recommendations?

    Thats ok.
    One of the good things of listening to this random radio, is that you listen to music without any form of prejudice. Some guys think that some genre is very bad without even know how this genre really is, and dont try this genre because of that. Some would even like this genre, but dont listen to this genre because some prejudice they have about this genre.

    Also sometimes you just want to try new things, and expand your taste, using this random station you would be able to do that.

    There also some times that you are just lazy, so you would just click on this radio and discover new things in a easy way.

    Redigerad av exdeathbr den 19 nov 2010, 16:15
    • belenen sa...
    • Användare
    • 22 sep 2010, 22:02

    "listening trends"

    I'd like the option to see the graph over my entire scrobbling history! also more color theme options would be lovely, like rainbow.

  • Random Radio an Extremely Valuable Idea!

    exdeathbr 's (post of 22 Sep 2010 11:10) regarding Random Radio represents not only a good idea, I think it is an important one; very self-actualization supportive (e.g., as in Michel Foucault and other individualistic freedom and rights-intensive thinkers who have famously articulated their concerns about sustaining the human 'self' -- quite enlightened, really, and markedly in contrast to some misinformed Orwellian or Malthus Book devotee-statistician as in "what Jane Doe 'should' be listening to or when John Doe 'should' retire [as in certain posts here from 2008. In John Cage's fascinating composition "Indeterminacy" he relates Isamu Noguchi's trenchant comment "Where is the 'Should' ] ).

    Random Radio looks to me as a friendly way to think about the "best and highest" uses of the superb technologies and their positive ethical potential via this vital last.fm. When I was responsible for making all of the a huge university's Music Dept.'s (now a fully accredited school of music) study tapes in its music library, for all of the myriad of the Dept.'s classes, I was delighted, as a professional musician and theory major + performance double, by the discoveries I --could not avoid-- making.

    From Tibetan monastery Mus. of Ethnomusicology Prof. 'K' to the Advanced Theory study tapes needs by Pulitzer Prize winning composer/Prof. 'A' and Music History Prof. 'L'. All this going on while, in the evenings performing blues, jazz, R&B, and Broadway show, ballroom and other 'easy listening' Pop music, or during the days, music for and in The Church [where Serious Art Music (in too many cases referred to by an instance-specific malapropism 'Classical' [that is: Serious Art Music written between 1750 - 1820]) began and grew up, from before renaissance genius Josquin Després -- try him! -- and, bet on it, last.fm does its typically great job of finding and very smoothly well-presenting him and, with that kernel then, the truly great music of his wonderfully talented contemporaries.]

    Because of all the practical necessity of my in-real-time (something functioning similarly to) Random Radio-esque "living conditions" I've described, now (the one and only individual self who is) "I" [Note: predicate nominative required; "me" is not a grammatical option, individual self-actualization is not a good use or rational excuse for tearing down the door-opening foundations of language, etc. : ) ], as a composer, a person, a Lifelong Learner and polymath, as a result I have a much enriched view of the blessings in this world and of what life on Earth has to offer, and which avails a vast spectrum-wide treasury of continued discovery and achievement (on a good day... an A.M. cup of the stuff I make that passes for 'coffee' helps. [Note: I can learn to make better coffee, too...]) Straight thinking, exdeathbr ! Nice job! -- esperancesp-com [This all relates the urgency to understand, as you clearly do, exdeathbr: re. trends and group "should" statistics applied -- or enforced! -- NLP-ed? Hideous. Horrific. "unspeakable mutilations of the spirit." -- W.S. Burroughs] "The Crowd is the un-truth."

    "You must be the change you wish to see in the world." --Mahatma Gandhi
    Addendum: [Change your thoughts and you change your world. --_______]
  • I like to see that you liked my text.
    You should look at one of my other text: The last journal text

    This text explain in a super easy way, with a analogy why most of population nowadays dont listen what they problably like to listen. You problably will like the text to.

    The random station idea was inspired in some way by this text on my journal.

    Last.fm recommendation is based on what people that listen to one artist or song, listen. The problem is that a extreme % of the last.fm userbase (mainstream listeners) have exposure with only a extreme minimal amount of artists and music. They depend on what tv and normal radio shows them (a extreme minimal amount of music).

    So, last.fm recommendations of mainstream artists are based on what those users know, because a person can not listen what they not know that exist. An example, if most of the userbase that listen to Rammstein didnt knew of any other band with industrial influences, no bands with industrial influences, would appear on recommendations. No matter if the users would like to listen to that.
    With the randomness of the my radio idea, the users try a huge amount of different things and the recommendation of this radio will not be limited by the lack of music exposure of many users.

    Another good idea about my station, is that in my random station idea, the songs an user dont like, influence the station recommendations too.
    There are many lady gaga listeners, that dont like britney spears, but this doenst influence the recommendation.
    In the same way that when someone likes likes alot ke$ha and likes lady gaga, ke$sha should be on lady gaga recommendation. If someone likes lady gaga, but hates britney spears, this should mean something to the recommendation.

  • But wat would realy want to see on last.fm (not only on playground) is a artist recommendation system that:
    1-use last 3 months data.
    2-use the most listened songs to get the artist recomendation. Last.fm would check the song and would see "what artists the guys that listen to this song listen".
    3-somehow make what I dont like influence what will be recommended to me
    4-Users with a higher amount of unique artists inlfuence more the recommendation.

    The number 1 is a more reliable chart to find what people really like. If you listen to a artist alot and find many better ones, it will take time to get them into the top overall chart.

    The number 2 would fix many problems, some problems with different artists with same name, the problem when you like just one song form a artist (but like alot this song) and last.fm recommend based on the entire artist.
    PS: I am not talking about doing like some recommendations websites (http://lastfm.heathaze.org/randomSongStats.aspx) that get recommended songs and group them by artists that made the songs and then recommend those artists, or just get the name of the artist that made the recommended song instead of getting the song. I am talking about really getting the artists that someone that like a song listen.

    The number 3, I dont know how this would work, but its a very good idea. This idea would help to fix a big problem with the recommendations to mainstream bands. I will explain with a example.

    Lets get nickelback:
    Imagine the guys that like nickelback. Now imagine that the chance of a nickelback fan that knows thar iron maiden exist also like this metal band is 5% (a very small % since they are different things). If 1000000 nickelback fans knows iron maiden, 50000 nickelback fans would also like iron maiden.

    Now imagine some underground band similar to nickelback called XXXX. Imagine that the chance that someone (that likes nickelback and know this) has to like this band is 80%
    The chance of some nickelback fan liking this band is 16 times higher then the iron maiden one.
    But the thing is: he would need to know that this band XXXX exist.
    The amount of guys that likes nickelback and knows that the band XXXX exist, would need to be 62500 to this band make a higher score in recommendation. Few underground bands has this amount of guys that know them. And so this artist will not be on recommendation even if they are very similar to nickelback.

    On this example if what you dont like, had some influence on the recommendations, those guys that like nickelback but doenst like iron maiden would influence the recommendation. Helping this underground band to get in recommendation charts.

    If the example, the chance of nickelback fan that know iron maiden hating, iron maiden is 20%. This amount would make the amount of hate become higher than the amount of like, making this band out of the list of recommendation bands.

    The number 4 is a good idea because users with more unique artists are more sure about what they listen because they experienced more things, and so not just think that his artists are good because of his lack of musical experiences.

  • Rating song while listening is good idea instead of loving track sometimes you font love 2 tracks same but still you can rate them 4 and 5 that doesnt mean you dont love first one but you love second more!

Anonyma användare kan inte skriva inlägg. Vänligen logga in eller skapa ett konto för att göra inlägg i forumen.