Designing APIs for unstructured data.

[Joyce Stack]

Hey Guys, 

We have lots of unstructured data that we cook to produce new datasets based on product requirements. We have to crunch lots of data which may get refreshed daily or weekly based on the particular data. These are all read only APIs. 

Examples of this are: 

• articles viewed by country 
• [{"views_by_country":[{"country_iso_code":"US","views":1,"views_last_month":1}]}
• co authors publication count
  
• [{"authorId":"57014531300","count":1},{"authorId":"57033680500","count":1},{"authorId":"8303189700","count":1}]

These APIs are really tied to the datasets underneath. We have a CKAN instance but we are also making the data accessible via APIs. 

I'm looking for any blog posts, books, exemplars in best practices in designing these sort of APIs that fall outside the CRUD realm.  

Cheers. 

[Mateusz Loskot]

On 1 February 2017 at 11:59, Joyce Stack <joyce...@gmail.com> wrote:
> I'm looking for any blog posts, books, exemplars in best practices in
> designing these sort of APIs that fall outside the CRUD realm.

Regarding the "APIs for unstructured data", I've been trying to research similar
problem and still learning...
You may find useful the questions I posed here and in other places.

"Modelling variety of attributes for resources"
https://groups.google.com/forum/#!topic/api-craft/geUAKGa1EuY
"Modelling data type with variety of attributes"

http://forums.raml.org/t/modelling-data-type-with-variety-of-attributes/1485

You may also find traces of some discussions in the archives of
the HTTP APIs; Slack room http://slack.httpapis.com/

Best regards,
--
Mateusz Loskot, http://mateusz.loskot.net

[Joyce Stack]

Thank you. 

I'll take a look.