Infrastructure workshop #opendatach 2017


#1

Slides and reactions from the Open Data Infrastructure workshop at the Opendata.ch/2017 conference.

In my introductory talk subtitled unconditional,basic,data, I used the metaphor of public water sources (springs and fountains, for which Switzerland and the town of Bern in particular, is famous for), to talk about some of the ways we can intuitively and objectively differentiate between various sources of data, and their downstream “reprocessed” variants.

I tried to paint a picture of a post-openness future, where we no longer have to explain to everyone the need for ubiquitous, raw, high quality data as something necessary for survival and prosperity. To try to bring the whole discussion to another level - of public data as another universal human/machine right. I’m glad that my presentation seemed to have had a motivating effect (i.e. not everyone fell asleep listening to my water meditation ;-)) - though, in retrospect, I could have spent more time demoing specific basic infrastructure projects (Data packages, Good Tables, OpenRefine, DAT, …) at the heart of our community activism and School of Data workshops. Links are sprinkled over my slides.

Luckily, my colleague in the association Andreas Amsler took care of doing just that - going into the details of how the CKAN APIs work, how open data catalogues can be synced together to create alternative, customized search engines for information - and not just for developers, citing a neat chatbot research project (PDF) from Austria. My takeaway from his talk is to keep working on projects like ckan embed and our prototype chatbot, and to try to raise interest in the community both for launching more portals - and for actively federating them together.

Andreas suggested that a shift in focus from data platforms to data services is going on among larger organisations, which are starting to see tangible efficiency and innovativeness gains in being more aware of their own data, and data in their field. He also suggested that we all try to click the links in his slides and play around with the portal’s API. I suggest pasting the links into this online JSON viewer to make that a little easier on the eyes.

Next, Barnaby Skinner, a friend and supporter of the community, stepped up next to show the gory technical and political details of his latest data journalism efforts. He already astounded me a few months ago with his progress, going from zero to hero with his Python skills. It’s clear that he’s not about to start slowing down. It was a very warm and fuzzy moment for me to fire up his Jupyter notebooks and see thousands of data points get scraped and analysed while we watched.

One particularly poignant (to me) point he made, was that opening data is not as interesting or important…or possibly as legal…as opening the code and process. I’ll need a while to let that sink in. Meanwhile, there lots more data journalism to explore in his slides and GitHub repos.

Finally, Juan Pablo Lovato pulled us together in the last talk, talking about the things he is doing to enable various departments and people in Swiss government to be a good partner to the open data community. He talked about the costs and efforts of creating the opendata.swiss platform so far, the kind of metrics (data inventory quantity vs. quality) and signs the various groups involved in OGD look at to judge progress.

Since Juan Pablo and I have been working closely together on projects like handbook.opendata.swiss for years now, I was as keen as he was to see the reaction from the group assembled at our workshop. He raised some really good questions to the audience, questions that - together with the critiques from his colleague raised earlier - waypost the way forward to better community platforms in the future:

  • how distant do we want these worlds (OGD, activists, business…) to really be?
  • and how should we close the publisher - developer - end-user loop?
  • what is the overall role of government and community?

Sehr spannend! (“very exciting” in German). Many thanks to the entire panel and our encouraging audience for taking part in the workshop, I really look forward to hearing your thoughts and feedback after the final keynotes.


[30.6] Python/Django meetup
Open data module at BFH
Data in the classroom