Creating linked data - shared learning from the process

I was recently tasked with converting the Global Hunger Index (not a govt dataset I know, but similar to some) into Linked Data - selecting how to model it and then modelling and publishing it online. 

As it was a learning project - the process and choices made are all documented online here (I used Google Refine to create SCOVO and RDF Data Cube versions of the data) in case they are of use to others.

We're going to continue to develop the dataset, adding in extra elements such as provenance information and discovery information. a

All comments / questions welcome.