Is it possible to write a provider like F # for related data?

Question

I really like providers like Freebase and World Bank, and I would like to learn more about type providers by writing one myself. A data program has been launched in the European Union where you can access data through SPARQL / Linked data. Would it be possible to wrap access to data to open EU data using a type provider or would it be a waste of time trying to figure out how to do this?

Access to EU data is described here: http://open-data.europa.eu/en/linked-data

Answer 1

I think this is possible - I talked with some people who are really interested in this (and are working on it, but I'm not sure what the current status is). In any case, I definitely believe that this is such a wide area that additional efforts will not be a waste of time.

The key issue when writing a type provider for data like RDF is choosing what should be considered as types (what should become the type name or property name) and what should be left as a value (returned as a list or key-value pair). This is quite obvious to WorldBank - the names of countries and properties become types (property names), and the values ​​become data. But for a triple-based dataset, this is less obvious.

So far, I think there are two approaches:

• Additional ontology - requires that the data source comes with some additional ontology, which indicates which keys to navigate. There is something called a “facet ontology” that is used by http://mspace.fm , and this can be quite interesting.

• Parameterization - parameterizes the type provider (in some way) and gives it a list of relationships that should become available at the type level (and you probably also need to provide some root where to start).

There are certain possibilities - and I think that a related data provider would be really interesting. If you want to do this for F # Data, there is a useful page for contributing :-) .