The FutureEverything conference today has been a veritable festival of open data talks and inputs, including inputs from Will Perrin, Chris Taggart, James Darling, Emer Coleman, Carole Goble, Nigel Shadbolt, Zach Beauvais, Jordan Hatcher and David Eaves. Whilst many speakers have been talking about the challenges of getting open data up online – challenging organisational culture and developing technical infrastructures, I’ve been trying to find a framework for thinking about the uses to which open government data is put. If the release of open government data is, in part, justified by the uses to which it can be put, then understanding those different uses can become important for thinking about ways in which data is released and how use is encouraged.
A bit of sketching in the sunshine outside the Contact Theatre Bar resulted in the following draft model:
To try and elaborate.
Services and Tools are uses of open government data that involve building something.
For example, someone might build a ‘find your nearest railway station‘ application, or a platform like Plings might provide information on Positive Activities for Young People taken from the Family Information Directory.
Exploration, engagement and advocacy captures uses where someone is exploring or using the data not in order to build something, but in order to engage with a particular issue.
For example, citizens reviewing charitable donation records, or a journalist looking at details of tube trains that passed red signals. As the result of these types of use an open government data user may: simply have a better understanding of an issue; campaign for something to be changed; highlight and suggest corrections to an error in government data; question government on a particular issue, and so-on.
Within each of these major categories of use, I’m suggesting a sub-division can be drawn, essentially between ‘sensitive’ and ‘non-sensitive’ uses of data.
Services and Tools
- Failure tolerant service and tool uses of open data are cases where inaccuracy in the data is unlikely to cause any significant harm or breach any duties of care a public body may have. If an application fails to update to the latest data, or mis-represents data, users simply have the choice of switching to another application.
- Failure sensitive service and tool uses of open data are cases where inaccuracy or misrepresentation of the data could cause significant harm, or conflict with the stringent duties of care many public bodies are under. For example, incorrectly advertising the times of activities for young people so that they travel to venues where activities are not taking place, potentially late at night; or having applications display out-of-date information about Accident and Emergency services could be a problem. (N.B: I know paper based information suffers many of the same problems of out of date data etc., so I’m not convinced these are the best examples – but relatively certain there are failure sensitive instances of open government data where inaccuracy / uncritical sharing of data risks causing harm)
Exploration, engagement and advocacy
- Collaborative uses of open government data for exploration, engagement and advocacy are cases where government is happy to work with the public in exploring and addressing and issue;
- Contested / Sensitive uses of open government data for exploration, engagement and advocacy are cases where the use citizens may make of data risks causing political upset and challenges to government. (Emer Coleman highlighted that often the political upset is a result of the culture from which government responds to problems being pointed out, and other speakers mentioned a ‘Daily Mail’ press culture as problematic here. )
To be clear: I do not think this draft typology provides any substantive arguments against the release of open government data, and it is not a critique of opening up access to data. Rather, I hope it can provide some foundations for thinking about which uses can be easily made of data, and which may be more complex and contested: and for reflecting on any issues this gives rise to for the tool-chain, infrastructure and processes of opening access to data. It’s also not yet a tested typology – so over the next few weeks as I build a dataset of embedded cases for the Open Data Impacts research I need to see whether this really does capture the range of uses of open government data taking place – or if some other set of categories is needed.
I’ll be around at FutureEverything tomorrow if anyone wants to explore this more; and blog comment space below for online exploration…
(MetaMethods node: This post is an experiment in participative research of online phenomena using this blog: not entirely sure that releasing ideas at such an early stage of formation is the best thing… but in the spirit of openness and just seeing what happens…)
Recent Comments