One of the deliverable features of this project is an integration with the CKAN.
To implement this the harvester framework for CKAN was installed into an instance of CKAN and then the data.bris havester plugin has been used to harvest the location and the META data of datasets.
From the systems point of view all that is required of it is for a RDF file to be generated which lists the META data for that dataset. This META data can include a set of tags.
It's these tags which are being used to allow for the discovery of other datasets. If the system generates the same tags for two different datasets they will be related and hence these tags will facilitate discovery.
The following are a few screenshots of the system in place. Currently they are showing the tags for columns in a dataset but it'll show the dataset level tags in the same way.
In future imports the automatically generated tags will have some sort of text appended to them to show that they were generated by this system and so would allow them to co-exist with user curated tags.