From getting people offline to providing more detail about formatting – the notes below were sent to me and other attendees of the Culture Hack North event today. I thought the points made about what developers want from data sets would be hugely useful to any organisation looking to get started with opening up data.
Thanking Ashley Mann of Opera North for agreeing to post it here and share it more widely – he’s @biglittlethings on twitter and well worth a follow!
Data:
As you have embarked down the road of thinking about what data you might release to the public I thought it might be helpful to share some feedback from the developers about what is/isn’t useful data, what formats are useable etc.
More is more: the more data you can release, the better!
Live: The developers were very keen to point out that the best data is live, it is something that they can access and be certain that it is completely up to date, there are a number of ways of achieving this that I’d be happy to explain if you are interested.
Accessible: I would like to stress that Word documents/pdfs are not useable formats as far as developers are concerned, csv or similar is best. Or simply (as the National Railway Museum have done) stick it into a Google doc spreadsheet – accessible and stored in the cloud so you don’t need to worry about bandwidth/servers/etc. A lot of the developers have said it’d be great if the data you provided for the event could be made available online.
Consistent: quite a few of you provided data that was quite inconsistent, for example dates that either referred to a specific date e.g. 1/1/2001 and/or a date range e.g. 1999-2005 in the same field – this renders this field completely useless to a developer or means that they have to go through and convert everything so that it is consistent (time-consuming and not a lot of fun)
Meaningful: some data was provided with no date or way of relating it to a specific time/place in the real world (i.e. a date), this makes it fairly unuseable. key pieces of information are place and time wherever possible
Semantic: Quite a few of the developers said it would’ve been nice to have more ‘human’ data, opinions, feedback, reviews etc. This would’ve allowed them to think more creatively about the use of the data.
Media: Developers also said that the best data was a mixture of figures, copy and also media e.g. photos and/or video
Engagement:
I think the organisations that probably got the most out of the event were those that managed to come along and be there in person. This not only allowed them to meet the developers but through doing so meant that they could explain their data in more depth and discuss potential ideas with the developers. I’d encourage you all to try and get along to any future events – I think there is a huge amount you could get out of it.
Dim Sum Digital 