Designing Open and Linked data apps is not easy
Over the last two years there has been much excitement about the idea of Open and Linked Data. This is especially so in countries like the UK where there has been a pronounced policy commitment to opening the use of public data for commercial and non commercial use. The UK government open data store boasts links to over 5400 sets of data saying “This can then be used by people to build useful applications that help society, or investigate how effective policy changes have been over time.”
There is no doubt that this data is of immense value to researchers. But despite various hack days, the number of genuinely useful applications seem limited.
We have been working with the data for the last nine months attempting to use labour market data to assist careers professionals and young people in choosing careers pathways. As Leia says in a comment on a recent post on this site “so many of our learners arrive with a complete incorrect (or no) idea about what skills are in demand and what’s realistic to expect in terms of looking for work and training.” We are not saying that labour market data and skills demand alone should guide young peoples’ choices. But it is certainly an important factor especially with university education becoming so expensive.
Why are we finding it hard to do? Firstly as the similar Salami project run by the University of Nottingham noted in a recent report much of the official data is collected for economic purposes, not for social use. For instance, much of the labour market information is collected through the Standard Industrial Classification (SIC) which although useful for analysing economic trends, is of limited use for occupational guidance. Instead, we really need Standard Occupational Classification (SOC) data. It doesn’t help that despite the data store which provides links to different data sets, the raw and interpreted data is scattered across a number of different web sites. Most of them are in the course of updating their sites, probably in order to make the data more accessible. But at the same time his is breaking links. And although there are a growing number of on-line tools, these all have their own idiosyncratic interfaces and processes (and often seem just not to work).
I was never very interested in statistics until I got involved in this project. And now I am desperately trying to teach myself SPSS but it is not easy and once more is time consuming.
Even when we have obtained the data it has to be cleaned. much of the data also requires manipulation if it is to be visualised. Visualisation tools are becoming more powerful, but still are not always simple to handle.
Using Open and Linked Data is a design process. And some of the most important people who have to be involved in any design process are the end users. Once more this is time consuming. And of course it is necessary to show them what the possibilities are. each different group of users will have different needs. We have spent a long time thinking about what data we should show to young people and what might be relevant for careers advisers.
Finally we have to remember that data is just data – however well visualised. The use of data has to involve meaning making. meaning making is not a precise science. Different people will make different meanings from the same data. The real added value comes when we allow them to participate in collective sense making through sharing and negotiating meanings.
We have developed the idea of a Technology Enhanced Boundary Object which is able to bring together data and data vidsualisations with a social software layer to explore meanings. We hope to pilot this in the autumn. And we will provide access to a working version of some of our tools in the next week.
So in conclusion – I remain very excited about the potential of Open and Linked Data. But to design apps which are useful takes a lot of work.
