Seasons started out as an experiment with the Spotify API, and a desire to do something which fills a gap in the typical radio format market. You can read the original announcement on my blog.
Essentially, the concept behind Seasons is a playlist with tracks that would have been played on popular music radio at this time of year over the past decades. So if you're listening to it in July, you'll hear songs that were on the radio in summer, and if you listen in January, you'll hear songs which were on the radio on winter. And so on, through the year.
Having created the initial playlist, and written the code to keep it updated on a daily basis, I realised it would be relatively simple to extend the concept to use a more restricted set of songs, in order to give a playlist which may be a little more listenable than the full fat version!
So, there are now five versions of the playlist, each with their own characteristics.
The original. A selection of songs from any time between the very first UK singles chart in 1952 and this time last year. You want eclectic? You've got eclectic!
A selection of songs that would have been on popular music radio in the heyday of AM broadcasting, starting with the pirate radio stations such as Radio Caroline and Radio London in the mid 1960s, along with Radio Luxembourg's English language service, and then running through early Radio 1 until FM took over in the late 1980s.
A selection of songs that would have been on popular music radio in the 70s, 80s and 90s. This is the basic format of existing "oldies" stations such as Greatest Hits and Gold Radio, but with the added bonus of getting a playlist that changes with the seasons (and without DJ chatter!).
The songs that would have been on the radio on this day in the two decades either side of the millennium.
A slightly different option, instead of being based on particular years and decades this covers the whole of the data, but only picking songs that are broadly classified as rock, alternative or indie. Note that this is even more experimental, at the moment, than the other playlists as genres are a bit harder to accurately get from the underlying data.
If you're interested, you can read more about how it works.