December 1, 2010 by nicolahughes
I recently attended an Open Data Master Class and I would like to share my thoughts, not as an expert but as a novice looking in on the CAR/Hacks Hackers/data journalism embryo. This is really a reflection on the nuggets of advice offered by some chieftains in the global village of data miners.
Open is suddenly cool - reflections on words by Dr. Hanif Rahemtulla (Nottingham University):
Data as it stands is not freely available; it's not truly open. Because only the people who know where to find it, how to use it and how to visualize it truly have access to data. Raw numbers are useless. If you don't understand the nature of it you can't mediate it. Yet there is an ongoing movement towards open data; in the UK, US, Canada, Australia, New Zealand, Ireland, Norway and even Kenya. 'Open' is suddenly cool because of open source
Linking is the future of open data - reflections on words by James Forrester (Data.gov.uk)
Data can be made pretty now - think Information is Beautiful. But we need to create many pliable tools so that we can customize the view to the viewer and to the story itself (note I mean story and not data). Data needs to be editorialized. You can pull a lot of things out of data that aren't true or worthwhile. But it's detailed data that enables data miners to go further. data needs to come with meta data
Making a web of data - reflections on words by Tim Hodson (Talis)
The simple fact of the matter is more tasks can be done by machines so making a web of data means not creating a different web or a new web but making the web we have better - scraping and linking. So once data is open the government needs to make it machine readable. Those interested in data mining - revealing the gem - need to provide the context by linking. But merging data bases is not fun so there's not many people will to go down the mines.
Mind mapping - reflections on words by Chris Parker and Ian Holt (Ordnance Survey)
Publication is not the same as communication. Nearly all data is published but very little gets communicated. One of the simplest (I did it in a day) and best tools for communicating data is to map it.