OpenBritain launch

For the past few months we've been busy at Osmosoft Towers working on an exciting project with with the folks at Open Britain.  Open Britain make available accessibility information associated to points of interest around the country.  Yesterday there was a 'soft' launch of the project at the House of Commons.  Unfortunately being at SXSW I was unable to attend, but others from Osmosoft were there in force mixing it up with ministers and the like.  There were even rumours of a sign-language flash mob involving David Cameron!

Anyway, here's the background to the project...

It started around a year ago with modest beginnings of providing some consultancy around the benefits of open data.  Open Britain, along with their partners, hold a number of datasets which up until now had been locked away in the depths somewhere.  Being advocates of open source and open data our view at Osmosoft was to make the data available in such a way that developers could build upon those datasets.  The more the data was being used the more valuable it'd become.

Back in May 2010 to help prove what might be possible we ran a three day hack event where four teams consisting of Osmosoft developers, data owners and individuals with accessibility needs were armed with a subset of accessibility data and brief to 'go build some cool stuff'!  Coming out of the hack event there was strong sense of excitement as to what might be possible should the data be made open.  However there were also some concerns around commercial sensitivities.

Over the following few months nothing was actually built with code however progress was made to mitigate the commercial concerns, and all parties involved agreed to a project to be managed by Osmosoft whereby data would be made available for non-commercial use.

It was agreed that Osmosoft would deliver three things: 

  1. an update to the Open Britain website
  2. a platform for capturing datasets
  3. a platform for developers to access and use the datasets

Work (and by that I mean coding) began again in December and the space of a few short months I'm proud of what we've produced as a team.

Update to the Open Britain website:

The original website hadn't been updated for a number of years and at a first glance you're hard pushed to know what purpose it really serves.  The homepage contains a number of search boxes, numerous adverts and links galore.

Our approach was to drastically simplify the user experience.  Discussions with Open Britain identified that the site had one primary goal and that was to provide visitors to the site with an easy way of searching the accessibility data.  There were also some contractual implications around advertising deals, but hopefully we managed to cater for these without implicating the user experience too much.

The site uses latest web techniques to aid the user as much as possible, such as using the browser to detect current location.  On completing the search the site retrieves matching results from the datastore in the backend (more on that in a moment).

The search results page has also been greatly simplified and provides the information that you'd expect from this type of site.

There is still some work to be done, navigation back to the search needs to be improved and there also needs to be some attention paid to the filters used.  For the moment we've taken a mixture of logical filters and ones that exist in the datasets.

I should also point out that the majority of the icons came from the fabulous noun project.

The data platform:

For us at Osmosoft the data platform was the exciting area.  It provided many areas of learning for us and also gave an opportunity to extend the TiddlySpace platform's geo capabilities.

For me one of the key areas of learning was around how to talk about what we wanted to do in such a way that it didn't scare the project participants from outside of Osmosoft!  It sounds like such a simple thing but when you're someone who uses the Web throughout the majority of the day it's easy to forget some of the preconceptions that others who are less reliant on the Web may have.  For this project a key area of concern was around crowd-sourced data.  Where we at Osmosoft saw it as potential gold mine of opportunities others in the project saw it as a pitfall of poor quality, unreliable data and potentially malicious risk.  It's a topic that has been debated for a long time and I'm pleased to say that we reached a sensible compromise. In the data sets we're indicating the provenance of the data and those who wish to only use data from 'trusted' data sources can simply choose not utilise the crowd-sourced data.

The data platform can be accessed from is built upon TiddlySpace platform and makes use of the platforms APIs to allow developers to use the data sets in their website, applications and mashups.  At the moment documentation is a bit thin on the ground but this will be improved upon over the coming weeks with full details of how to use the APIs and how to add datasets.  We'll also be adding to the datasets from a number of sources and providing a mobile web-app to allow individual point of interest data to be captured.

I'm looking forward to getting back to London to find out how the launch went and begin working on the next steps to make the open data platform a success.


Guardian SXSW Hack Weekend 2011

Over the weekend of February 12th/13th The Guardian hosted a two day hackathon.  The purpose of which was to consider the obstacles faced by journalists and reporters whilst covering events such as SXSW.  An alternate, slightly broader objective was to simply build "cool shit".  Jemima Kiss from The Guardian explains more about the event here.

In the build up to the hack weekend I'd been thinking about the kind of problem I'd like to tackle.  A problem which seemed to be particularly relevant for SXSW due to the sheer number bands playing, 1,900+ over 5 days, was figuring out who to go and see.  Looking at the SXSW schedule is a daunting task to find out about bands you've never heard of – and lets face it, out of 1,900 bands there gonna be quite a few who are new to you.

Since picking up my ticket for SXSW I've been busy finding out about who's playing.  This has typically consisted of a lot time on LastFM, creating a Spotify playlist, reading up on Wikipedia entries, and generally soaking up what the web has to offer.  All of which has been fun, but not something that I want to do for all 1,900 bands.

So coming into the hack weekend I thought it might be cool to have something which listed all the bands playing at a festival and against each one pulled in information associated to the band from across the web displaying it all in one place.  Combining this information with the location, date and time of the gig would then allow me to build up a schedule of who I wanted to see.  I also wanted someway of recording my notes about the band that could then be shared with others.

All of this sounds pretty doable, but not being a programmer I wasn't sure where to start!  Thankfully I was able to buddy up with Ben Nickolls and Paul Tanner who were both interested in tackling the same problem.

Over the course of the two days what we were building changed approach a couple times due to reasons partially technical, partially through time constraints and partially through my realisation that trying to learn (basic) JavaScript at 4am wasn't cool.

What we ended up with at the end of the weekend was a two-pronged solution.  A webapp built using jQTouch that uses data provided from both Robbie Clutton's and Matt Andrews's APIs to display a list of all the bands playing at SXSW and then a breakdown of a suggested schedule based upon your LastFM listening habits (a massive thank you to Robbie and Matt for providing the APIs).  From the webapp you can see some basic information about the band and if you wanted more you could then click through to a more detailed view where you can also add your own notes to be shared with others.

The detailed view of the band we built using TiddlySpace where with the help of Paul Downey's LastFM plugin it pulls in information relating to that band.  It still needs a bit of love but we hope to continue working on it over the next few days, you can keep track of what we do here and at

Overall the hack weekend was incredible fun and although I left feeling exhausted I'm eager for the next one to come around.  I learnt a huge amount over the two days and made some great new friends.  Some awesome hacks were built and you can see the full list here.

Oh, and we also got so cool retro swag:

Retro schwag! #GSxSW



My top albums of 2010

Seeing as everyone else seems to be doing it I figured that I may as well put down my top 10 albums of 2010.  Here's the list in a rough ten through to one. No laughing please.

10 - Learning (Perfume Genius)

9 – I'm New Here (Gil ~Scott-Heron)


8 – Causers Of This (Toro Y Moi)

7 – So Runs the World Away (Josh Ritter)

6 – Sex With An X (The Vaselines)

5 – Age of Adz (Sufjan Stevens)

4 – Í annan heim (Rökkurró)

3 – High Violet (The National)

2 – Majesty Shredding (Superchunk)

1 – Mount Chimaera (Brasstronaut)