Feedback from Chris Orton:
I understand that there are changes on their way re metadata onboarding, but I would just like to note a few things in the meantime.
Having the ‘Observations’ section on the HOP as distinct to ‘Provenance’ is confusing when the fields based on the Observations appear on the Gateway under Provenance.
Given, for example the Wales Asthma Observatory, has no distinct observations per se, which I/we would see as specific analytical findings or properties of the dataset such as variable presence, the fields under Provenance on the Gateway ranging from ‘Statistical population’ to ‘Observation date’ remain blank.
I could easily provide the overall statistical population of this dataset (~190,000), the description (those with Asthma presence in Welsh health data), and the population size (population of Wales to create the cohort), but given this doesn’t appear under ‘Provenance’ in the HOP, but does on the Gateway, there is no obvious way of knowing this is how the categories link together. I would also say that an ‘Observation’ as many of our data providers have supplied to us, is based on a finding or specific variable of interest in the data, and fields such as ‘Population size’ etc. tend to be more static information which is relevant to the power and description of the base dataset used.
Could this be incorporated into a more sensible flow of information from HOP to Gateway, or when the onboarding function moves to the Gateway, be borne in mind?
Thanks for your feedback. Yes I completely agree that observations ought be in a separate category from Provenance. We had to make a pragmatic choice in the early days when Observations were not completed. Now that we are reaching a critical mass, it would make sense to separate this out.