The statistics should not be interpreted as the number of items in those repositories, but rather as the numbers of items (full-text items) that can be harvested from those repositories using OAI-PMH. Please also do note the dashboard is work in progress, so the stats might not be completely accurate yet.

There is a huge discrepancy in the way repositories expose metadata about their content through OAI-PMH, which dramatically influences the content harvestability. Taking into account only EPrints repositories (that are typically quite good in referencing full-texts) from the UK, the average repository will have for about 27.6% metadata records harvestable full-texts (but the median repository only about 13%).

The aim of Repository Analytics is to help repository managers to identify possible issues. The statistics can be also collected through the CORE API (http://core.kmi.open.ac.uk/api/doc – API stats methods), for those who do not like UIs. I have originally created some recommendations on the CORE website to increase the harvestability (http://core.kmi.open.ac.uk/intro/core_recommendations) of content from repositories. I have been in the last moth thinking about how they could be even more simplified. I just submitted a paper to OR 2013 about this. Happy to send it to those interested. Will make it publicly available if accepted :)

Realise you are probably aware but Leeds, York and Sheffield all come under the White Rose consortium of course – White Rose Research Online (WRRO) which doesn’t affect your basic point.

CORE is still being developed I think – try @petrknoth on twitter.