Reflections on Oxford Open Data Day

[Summary: creations and learning from Oxford Open Data Day]

Yesterday around 30 people got together in Oxford to take part in the first international Open Data Day, an initiative sparked off by David Eaves to get groups around the world exploring what they could create with public data. For many of the assembled Oxford crowd it was their first experience of both exploring public data, and taking part in a hack-day event, so, having started at 10am, it was fantastic that by 4.30pm we:

• Knew a lot more about where Arts Council Funding was spent across the Country, and specifically in the South East.
• Had learnt at lot about the availability (or lack thereof) of good data on taxation, and had created visualisations of variation in income across the UK, as well as creating the framework for a rather funky visualisation of what a tax on bankers bonuses could raise.
• Understood a whole lot more about the CKAN API and the meta-data available on data.gov.uk (which is turns out is pretty ropey right now, and could perhaps do with some data-librarian’s attention…)
• Discovered, via analysing the sentiment of Twitter messages that Margate appears to be happier than Oxford; or at least it’s Twitterati are;
• Had a working day of a fantastically useful live bus map for Oxford.
• And had a fantastic new Blue Plaque generator that will return interesting facts about nearby famous people wherever you are in the world using data from Wikipedia.

Thanks to everyone who took part in the day, and particularly to Ed, Kevin, Ed & Dave at White October for hosting the event, and to Incuna for sponsoring the lunch. Many thanks also to Sywia for blogging the event: you can find photos and video clips sharing the story here.

Quick Learning Notes

Skill building: I also took advantage of the Open Data Day to start exploring some of the ideas that might go into an Open Data Cook Book of ‘recipes’ for creating and working with open data. There are big challenges when it comes to building the capacity of both technical developers and non-developers alike to discover and then work with open data.

I’ve been reflecting on the discovery and design processes we could make use of at the start of any open data focussed workshops – whether with developers, civil servants, community groups or campaigners to provide the right level of context on what open data is, the potential and limitations of different datasets, and to provide a general awareness of where data can be discovered. At Open Data Day in Oxford we perhaps struggled to generate ideas for projects in the first half of the day – but understandably so given it takes a while to get familiar with the datasets available.

I wonder if for hack-day style events with people new to open data, some sort of training & team-building exercises for the first hour might be useful?

Data-led or problem-led: Most of the groups working were broadly data-led. They found some data of interest, and then explored what could be done with it. One group (the visualisations of impacts of tax changes for the Robin Hood Tax campaign) was more ‘problem led’ – starting with an issue to explore and then seeking data to work with. Both have their challenges: with the first, projects can struggle to find a focus; with the latter, it’s easy to get stuck because the data you imagine might be available turns out not to be. Finding the data you need isn’t available can provide a good spark for more open data campaigning (why, for example, are the details of prices in the Retail Price Index basket of goods not being published, and FOI requests for them being turned down on the basis of ‘personal information’ exemptions?), but when you can’t get that campaigning to produce results during the course of a single day, it can be pretty frustrating as well.

On the day or in advance?:
We held a pre-meeting for the Oxford Open Data Day – and it was useful in getting people to know each other and to discover some ideas and sources of data – but we perhaps didn’t carry through the ideas from that meeting into the hack-day very strongly. Encouraging a few more people to act as project leaders in advance may have been useful to for enabling those who came wanting to help on projects rather than create their own to get involved.

Data not just for developers:
My mantra. Yet still hard to plan for and make work. Perhaps trying to include a greater training element into a hack day would help here, or encouraging some technically-inclined folk to take on a role of data-facilitators – helping non-developers get the data into a shape they need for working with it in non-technical ways. Hopefully some of the open data cook book recipes might be useful here.

Sharing learning rather than simply products:
David Eaves set out three shared goals for the Open Data Day events:

1. Have fun

2. Help foster local supportive and diverse communities of people who advocate for open data

3. Help raise awareness of open data, why it matters, by building sites and applications

emphasising the importance of producing tangible things to demonstrate the potential of open data. This is definitely important – but I think we probably missed a trick by focussing on the products of the hack-day in presentations at the end of the day, rather than the learning and new skills people had picked up and could tell others about.