Ways to publish Open Data on the Web

[toka (Thomas Kalka)]

Lets continue the the discussion from Transition Network checking in about opening up data here:

Transition Network checking in

• We’d love to see this data presented in other places, particularly on the base layer of Open Street Map.
• Are there things we can do to make our Initiative location data more accessible to you? Is GeoRSS useful to you? http://en.wikipedia.org/wiki/GeoRSS

@SamRossiter
The first step in opening up data would be to provide some machine readable access method (api).

This could be done in different ways, which include more or less work on your side.

(A) open up read only access to the databases you use to create your website, limited to view public data only

(+) needs little work
(-) potential security issues, difficulty to implement this, depending on your database

(B) create views with public data in your databases, copy them regularly to a database with open read access

(+) more secure compared to (A), should be not so much work
(-) no built in versioning, so no easy updates on followers

© create a bidirectional data flow into some couchbase instance
(+) this seems already a near optimal solution, you get all features of couchdb for free: standardized web api, easy replication, changes feed
(-) more work than (A) and (B),

To implement this, you will need to introduce some concepts of couchdb into your database, like usage of uuids and version id’s for your entries.

(D) create a in page data representation of public data (some json, which can be parsed out of the page)
(+) can be implemented in the data transformation layer, which you use allready to render your site
(+) benefits everyone, who wants to scrape your data
(-) changes feed has to be implemented separately

(E) variant of (D) using http://microformats.org/
(-) more complicated than (D)
(-) you will change allready existing code to produce your website
(+) supports microformats

(F) there are possibly other emerging standards to include easy parsable data in web pages
We should collect them somewhere.

• microdata: http://diveintohtml5.info/extensibility.html
• https://en.wikipedia.org/wiki/ICBM_address

(G) in general we should find ways to publish and link to data / services inside the pages visible to humans

• there could be a link inside the meta section to a machine readable data representation (either json, yaml or xml)
• there could be a link inside the meta section to a changes feed for this type of items
• there could be a link inside the meta section to an directory service for this type of items

[SamRossiter]

Hi @toka thanks for this, I’ll ping @Annesley_Newholm who’d be responsible for implementing any changes as far as Transition Network are concerned…

Thanks

Sam

[Annesley_Newholm]

we will be creating an API. the schemas are under development already. at some point an XSLT will be written to provide others with public access.

Annesley

[toka (Thomas Kalka)]

Happy to hear that. Keep us informed.

Thomas

[Annesley_Newholm]

Transition Network is examining a completely de-centralised Semantic web approach to data where the end entities themselves (e.g. http://www.transitiontownbrixton.org/) publish in-page RDF data stating, amongst other things, their schema:Place and schema:Organisation (TransitionTowns).

TransitionTown entities (websites) could then be programmatically discovered through SPARQL web engine searches using something like Yahoo Glimmer:
http://glimmer.research.yahoo.com/#!index=wdc&query=type%3A{http%3A%2F%2Fschema.org%2FOrganisation}+Transition&deref=true&pageSize=10&pageStart=0

Thus, there would be no central database, website or API. No Transition Town would need to register in a central website anymore. They would, instead, install a plugin / module on their site which published all relevant RDF data on their homepage to be discovered by RDF capable Internet search engines. The same plugin would spread peer-to-peer by advertising itself to visitors offering itself for installation on the visitors website in order to also be considered a TransitionTown.

Debates, decisions, visits and information would then happen through the plugin via constant decentralised SPARQL queries. Entering a post in a debate would be the same as posting on your own blog with the correct rdf:category. Viewing a debate would be a SPARQL query. All this would be invisible to the end user.

GRDDL would be used to transform existing WordPress and Drupal schemas spaces.

The website admin interface would require extra categorisation of posts in some cases

Thoughts? Elaborations?.

[species (Michael Maier)]

Hi @Annesley_Newholm,

In the future, we also wanted to build upon RDF, als @almereyda will confirm.

Your vision sounds very bold and I would like to see it implemented in TransforMap too!

Have you thought about initiatives not owning a website where they can publish RDF data? Some groups maybe only have a Facebook page, or are completely offline - how to include them?

I would be very pleased to see you in our Monkey’s circle - are you planning to join us at our
Mapping Meetup Munich in June?

[Annesley_Newholm]

Yes, you are absolutely right. At the moment we envisage that a registration service would still be available for groups without their own websites. That registration service may be part of the RDF plugin on another Transition Town website of course, rather than a central authority.

Monkey’s circle sounds fun. we will see

[mattw (Matt Wallis)]

Transition Network is examining a completely de-centralised Semantic web approach to data where the end entities themselves (e.g. http://www.transitiontownbrixton.org/) publish in-page RDF data stating, amongst other things, their schema:Place and schema:Organisation (TransitionTowns).

Hi @Annesley_Newholm, do you have any findings from this examination. The Institute of Solidarity Economics have published the ISE Strategy for Data which is examining the same de-centralised RDF approach. There is a discussion in another thread on this discourse. It would be great to hear about your experience.