Many people in the University support the idea of making resources
open for anyone to use, and some trailblazers have set up the Open Knowledge Network to help support this notion. One aspect of this is open data and with the Edinburgh Cityscope
project developing nicely, it would be timely to put support for open
data on a firmer basis. In this post, I consider some implications and
prerequisites for publishing University data.
Suppose someone wanted to make some data available as open data. What would they need to consider?
Well, first of all, who owns the data, and who is responsible for it? If it is the University’s data, whose permission do they need? To answer this, we have agreed policy to assign data stewards to the University’s main enterprise data sets. These data stewards will be responsible for making data available to people who need it, and ensuring that restricted data (such as personal information) is protected. So they will be the people to give or deny permission.
A follow-up question is, what is the University’s policy towards making data open? As far as I'm aware, there is no official policy specifically about open data. The University does have a policy that empowers people to make educational resources (such as lecture notes) open, and it would seem sensible to clarify whether there is a similar institutional desire for publishing data. So a group of us are writing a draft policy to take to the official policy-making committees of the University.
The next question is whether the particular data in question is allowed to be published. While the University might encourage the idea of open data, there is plenty of data that we should not publish. This includes personal information, certain financial information, data provided in confidence, and so on. To help the data stewards, and University staff in general, we are proposing a simple scheme for classifying the confidentiality of any information into one of three categories. Only data in the first category – unrestricted – may be made open.
The Data Governance Group will give data stewards advice and guidance as to which data should be classified into which category. Enterprise Architecture will add this to our nascent data dictionary, so we have a central, shared record.
Then there are practicalities to consider. Which open licence should be used (e.g. Creative Commons)? Is the data sufficiently well documented? What processes are in place to handle queries, and to publish corrections? Will the data be kept up to date? For all these concerns (and more besides), we need to provide guidance to help people make consistent decisions.
Open data has been an ambition of the University for several years. Now it is looking likely we can put this ambition into practice.
Suppose someone wanted to make some data available as open data. What would they need to consider?
Well, first of all, who owns the data, and who is responsible for it? If it is the University’s data, whose permission do they need? To answer this, we have agreed policy to assign data stewards to the University’s main enterprise data sets. These data stewards will be responsible for making data available to people who need it, and ensuring that restricted data (such as personal information) is protected. So they will be the people to give or deny permission.
A follow-up question is, what is the University’s policy towards making data open? As far as I'm aware, there is no official policy specifically about open data. The University does have a policy that empowers people to make educational resources (such as lecture notes) open, and it would seem sensible to clarify whether there is a similar institutional desire for publishing data. So a group of us are writing a draft policy to take to the official policy-making committees of the University.
The next question is whether the particular data in question is allowed to be published. While the University might encourage the idea of open data, there is plenty of data that we should not publish. This includes personal information, certain financial information, data provided in confidence, and so on. To help the data stewards, and University staff in general, we are proposing a simple scheme for classifying the confidentiality of any information into one of three categories. Only data in the first category – unrestricted – may be made open.
The Data Governance Group will give data stewards advice and guidance as to which data should be classified into which category. Enterprise Architecture will add this to our nascent data dictionary, so we have a central, shared record.
Then there are practicalities to consider. Which open licence should be used (e.g. Creative Commons)? Is the data sufficiently well documented? What processes are in place to handle queries, and to publish corrections? Will the data be kept up to date? For all these concerns (and more besides), we need to provide guidance to help people make consistent decisions.
Open data has been an ambition of the University for several years. Now it is looking likely we can put this ambition into practice.
Comments