5 things you always wanted to know about multidimensional data and co-creation but were afraid to ask

Published in

opengovintelligence

4 min readNov 23, 2018

Me, pointing. Photo & tweet from Gianfranco Cecconi

Yesterday I went to Delft University for a co-creation and linked data conference, which was all kinds of good. I had a job whittling down my takeaways to five but here they are (& there are some links at the foot of the post for anyone interested in more about the day).

Challenges of publishing data

Open Data is open, not just when you publish it but when people can find and use it — Linda van den Brink

Linda van den Brink, Darren Barnes (Open Datasets Manager, ONS) and Gianfranco Cecconi (EU Data Portal) were our first panellists. They agreed that three of the main challenges of publishing data are:

findability
usability
consistent vocabulary — so you know when you find a term in one piece of data and in another that they are related regardless of the publisher.

Darren Barnes on the importance of data infrastructure (source: twitter @OpenGovInt)

Co-Creation

Co-creation isn’t about offering people a red or yellow box.It’s about asking if users even need a box. — Max Kortlander (of the Waag Society)

The co-creation part of the day was great — in part because of the energy of the speaker and panellists (take a bow Max Kortlander, Jamie Whyte and Marijn Janssen) but also because the message was clear: co-creation isn’t a passive process where a developer / organisation has an idea which solves a problem and users change the details. Instead, users contribute to the overall solution and explore possibilities together with those who can make it happen. Panellists discussed how timely communication motivates those users taking part in co-creation; that multiple iterations are key and that quality in co-creation doesn’t always mean more people involved but, rather, the right people being involved (those who truly represent the group the service is for).

Data is Messy: bring on the tools

Data is messy. What kind of tools ease the pain to get to good standards based, interoperable data? — Bill Roberts

We got a real explanation of the open source tools table2qb and CubiQL which are being used in the OpenGovIntelligence project and which you can find on github. These tools are about the developer experience rather than the end user, so they can produce tools easily. In brief, table2qb is a way of creating linked data cubes from tabular data. You can then point the CubiQL tool at it and it gives a developer friendly API. CubiQL makes it easier to understand what’s in the data and the API can be used to write applications.

Show the Thing

After all the talk of tools you want to see a thing — and by that I mean an example of end to end use of tools. Trafford Data Lab have developed three apps on the subject of worklessness as part of the OpenGovIntelligence project. Scan (above) identifies statistical hot and cold spots of worklessness clusters and does so by:

Collecting csv data of claimant counts
Generating rdf with the table2qb tool
Putting that rdf into a triplestore (database)
Pointing CubiQL at the database

When you click on an area on the map, the app calls CubiQL to get the calculations which colour the map. Another pilot using the tools is the Irish pilot (which has a renewable wave energy focus). The Marine Institute use table2qb to generate their data and a front end application that uses CubiQL to do the same thing.

Bringing people together

ODM and Waag: both rather fond of data and working towards a fairer society

A big part of what made yesterday so engaging was how much knowledge, experience and curiosity people had in common. Honestly, it was like Blind Date where they end up getting married. I personally have high hopes for