Box UK: BBC Innovation Labs Blog » travel

Box UK and the BBC – the beginning!

Claire S — Tue, 15 Jul 2008 16:20:07 +0000

Over the last week or two we’ve started work on the initial stages of our project with the BBC, which we’ll be keeping you up to date with over the next couple of months.

The initial stages will be very much R&D related and will involve experimenting with the BBC’s existing data. One of the issues that we’ll need to deal with is coming up with amazing, sexy but useful applications that users won’t be able to live without. Dan’s previous posts go to show some of the scope that can be covered with just travel data.

When we were picked to take part in the Labs we went in thinking that, by restricting ourselves to just looking at travel data , that we had a fairly limited scope, and I’m happy to admit that there was a slight feeling of jealousy at the companies sitting in the room with us talking about ARGs and interactive television shows.

However, when we sat down and started looking at the sheer amount and variety of travel data that’s available, the platforms and media that it can be pushed out on to, and the other data that can be mashed with it, we’ve come to realise that there is huge potential to develop applications that are sophisticated, fun, useful and visually stimulating… and we’re getting very excited about the whole thing!

So we’re going to regularly write about all the cool stuff we’re doing and if you have any thoughts or comments along the way please let us know…

*** Claire

Don’t drive to Cardiff on a Monday…

Dan Z — Tue, 27 May 2008 13:18:27 +0000

Our database of incidents has been growing over the last couple of weeks, since we started recording data from the BBC Travel feed.

The dataset size is still relatively small, for meaningful statistical analysis, but I just thought I’d perform some random mining of the data, for fun. (Yes, ‘fun’… I do need to get out more).

From initial results, it looks like Monday is the worst day for driving:

And, if you’re travelling along the M4, the city of Cardiff – although not the largest urban area along the length of the motorway – seems to suffer from the most incidents (clearly, it’s the large number of tourists flocking to our beautiful city).

More coming soon…

*** Dan

Personalised BBC Travel XML Feeds

Dan Z — Wed, 14 May 2008 14:27:45 +0000

OK, so I’ve updated my little experiment so that you can now request personalised travel XML data, which is taken and filtered from the original huge BBC travel XML feed.

The root URL for the feed is:

http://feeds.boxuk.com/bbctravel/xmlfeed/

You can use the following parameters to personalise the data:

road (either a road number or name)
place (this one’s a bit odd at the moment due to the original data, some cities work, some counties work)
severity (either very severe, severe, medium or slight)

Multiple values can be separated with commas, and the parameters can be specified in any order. Example feeds are:

I’m hoping to add in some longitude/latitude parameters soon, so that you can request a feed for just a particular geographic area.

If you have any ideas or comments, please add them to this post!

*** Dan

PS This is just a hack for the time being, with little in the way of caching or robustness… Please do not rely on these feeds!

UPDATE: Just added a little dynamic ‘summary’ page too from the live/historic data too, which I’ll try to add some more graphs to…

Using the BBC Feeds: Whetting the Appetite

Dan Z — Wed, 14 May 2008 09:48:42 +0000

I spent a few hours yesterday setting up a script that periodically grabs the BBC Road Travel XML data, parses it into a database, then grabs the current weather conditions for any new incidents (for the local area), and stores that alongside each incident. Hopefully in a week-or-two there’ll be enough data to produce some interesting graphs, such as number of incidents vs. weather conditions, or number of incidents vs. time of day.

In the meantime, just out of interest, I’ve quickly grabbed the BBC 7 Day Listings data, and grep’ed it for some different genres of output that the BBC is pushing to its audience in the next week. As a committed atheist with a scientific background, it’s a bit dis-heartening to see this kind of balance of output from the BBC:

With about half of British people now having no religious beliefs, and I expect a slightly higher percentage having some belief in Science, I wonder how long this situation can last before the public demands a more even balance.

*** Dan

Michael Fish knows why you were late…

Dan Z — Fri, 09 May 2008 14:50:01 +0000

Claire and I are keen to use the large and varied data sources of the BBC to provide unique, beneficial services to the end-user. One way we think we can do this is by combining and comparing (‘mashing up’) the multiple sources, where relevant, to produce extra insight.

One of the avenues we’re keen to explore is how travel and weather data can be combined: how weather can affect travel/driving conditions, and therefore how a weather forecast could predict potential problems.

An obvious example of this is extreme weather: snow, heavy rain and high winds will have a noticable impact on driving conditions. But, given the beautiful conditions recently, I’ve also been wondering how sun may also be potentially included on this list of conditions that could cause problems.

As a very quick (and un-scientific) test, I’ve been checking the BBC’s road alerts XML feed this afternoon (for about the last hour, since 3pm). Specifically, I’ve been looking for the number of incidents that are ‘westbound’ (facing the setting sun) rather than ‘eastbound’. There’s little statistical significance, but the number of westbound incidents have been consistently 10-15% higher.

I’d like to take this experiment a little further. If I have time, I’ll set up some scripts next week that pull-in and record the BBC’s weather and travel feeds, and log the data to a simple database. At the end of the week, I’ll then be able to compare and contrast various types of road incidents against the weather conditions at the time (be it rain, sun, or something else), and the time of day. Hopefully it should make for interesting analysis, and if it proves slightly significant, could be something we consider using at a more regional level as part of our work with the BBC.

*** Dan

PS For those of you tuning in to find out about the next Cardiff Web Scene meet-up, we’re finalising a venue, and should have more information early next week! Stay tuned…