People listen to songs all the time and usually have a fondness for certain specific types of songs, but they can never find out what the song sounds like without listening to it first. Also the genres of music are very loosely defined so they cannot really depend on the genre to decide if they are going to like the song or not. Also a person may want to listen to a specific type of song at a time depending on their mood. So what the project proposes to do is analyze the songs using their characteristics like frequency, beats etc to build a Self Organizing Map (SOM) which places similar types of songs together. These SOMs can then be used to build playlist of songs. These SOMs can also be uploaded to a central server. This central server will then accumulate SOMs of all the users and combine them to into a detailed SOM. This SOM can be accessed by users to find out about new songs that are similar to the songs they already like. Thus the program allows users to experience new songs that they most probably will like because they like similar songs.