The John Peel Time Machine

Posted on 04/10/2012 by


Guest Post by Storm’s Mike Ellis

Everyone knows the legend that is John Peel—and everyone who knows John Peel knows the Peel Sessions.

For 37 years, from 1967 to 2004, a vast range of bands recorded live in the BBC studio – normally laying down 4 tracks; some of them rough and ready, others polished and neat – but all of them resonating with energy and passion.

Peel’s particular specialism was actually his lack of specialism – the music he promoted and featured on the Sessions ranged from reggae to dance to pop to death metal. The bands he featured were often experiencing national coverage for the first time via his show – and – possibly as a consequence of this – often went on to follow a trajectory of enormous fame. The Sessions therefore map out a large and important chunk of music history – a bit like following a band from their first appearance at your local club to their appearance at Wembley five years down the line.

When you go to the Wikipedia page which features the entire list of the Peel Sessions, the first thing you notice is that there is lot of scrolling to do..

The size of this dataset sets the backdrop to the challenge that we’re trying to solve with our Kasabi-driven app, The John Peel Time Machine. We wanted to try and deliver an interface which delivers the richness of the Sessions – and all the potential associated data as well – without overwhelming the user.

Our basic user interface device is a draggable timeline, built using jQuery – but with a series of filters included so that users can either browse or dig in if they know what they’re looking for.

Clicking on a session pops up an overlay box with further information about the session, including an artist biography, track listing, links to Spotify or iTunes where relevant and artist photos:

The data is mostly derived from several different Kasabi datasets: the sessions come from the John Peel Archive (which is itself derived from the awesome John Peel Wiki), while the artist information is pulled from Discogs and the also awesome Dbpedia—a linked-data representation of some of Wikipedia.

One of the things we spend a lot of time doing at Storm Consultancy is understanding and using new technologies which couple cross-browser consistency with deep interactive capability. Once upon a time, this kind of functionality would have been built in Flash. Nowadays, the options provided by jQuery and html5 are becoming rich enough to support this kind of faceted, drag-drop approach – and also do it cross-browser, without the need for any proprietary plugins or downloads.

Some of the particular challenges have included…

  • Dealing with such a large dataset. There are 3,662 session, by 1,895 artists and groups, and 10,396 tracks, all with connected meta-data. Wrangling that much data on the server is easy, representing it in a browser is rather more of a challenge.
  • The incompleteness of the data, especially from the early days. Many of the bands were not around for long so we don’t know much about them, and record-keeping was often pretty incomplete.

This has been a fascinating project for us to undertake – we’re obsessed with the Sessions and the Kasabi data offering has really given us a unique opportunity to blend user experience design with a rich dataset to provide what we hope will be an interesting way of delving into a piece of music history.

The John Peel Timemachine is an ongoing experiment, and we’ll be adding more data, and more ways to navigate around the sessions over the next few months.

If you would like to get in touch with Talis learn more about working with data-rich applications, drop us a line.

Posted in: Funding, Ideas