For over six years, OUCS and then IT Services have provided OxPoints, a dataset and web service that catalogues the collegiate University’s organisational structure and physical estate.
As time has gone on, the limitations of the current implementation have become more apparent. First and foremost is that all its data has to be manually curated, even if there are perfectly good authoritative sources we could be calling upon. We could develop ways of synching in external data, but there’s no way of recording provenance, so we can’t say of a particular data item “this came from this upstream; don’t attempt to edit it here).
Closely related to the above issue, there’s no API for other people to push data in. Say, for example, that a department has a room booking management system that knows the capacities of rooms in various layouts. They can’t automatically update OxPoints whenever their data changes, creating either duplication of effort through rekeying, or stale data in OxPoints.
OxPoints also doesn’t support providing real-time data feeds of changes.
There are loads of external data sources we could be making use of, and which are far better maintained than our current dataset could ever be if it continues to stand on its own. Sources of data we’d love to integrate include:
- Planon (building, space and floorplan data for the University estate)
- Registration (organisational unit information for the collegiate University)
- LibraryData (library metadata from the Bodleian)
- OpenStreetMap (building outlines and metadata)
- OpenCharities (e.g. information from the Charity Commission about colleges)
- OpenCorporates (e.g. information from Companies House about the University’s corporate endeavours)
- DBpedia (linking out to provide more context to the information we hold)
The way forward
IT Services is in the early stages of a project to replace OxPoints with a new system that can better integrate data from multiple sources. We’ll be engaging key data owners, such as PRAS, Estates Services and IT Services’ IAM team. We’ll also seek requirements from potential consumers to encourage them not to maintain parallel overlapping datasets, and instead pull the data they need from this new system.
We hope to provide a system that can be easily extended to integrate new data sources as they emerge, and where that integration can be done bit-by-bit. Real-time data integration will also be key; the chaining of data imports with the current system means that it can take a while for data to be updated in e.g. maps.ox, but that it’s difficult to know when the data will finally make its way there.
So what might we expose? Things like:
- names, identifiers, addresses, locations, outlines, occupants, accessibility information, and images
- Names, ordering and semantic floorplan images
- Spaces and rooms
- names, identifiers, categorizations, occupants, layouts and capacities, images, features, and links to calendaring/availability sources
- Departments and units
- names, identifiers, sites, organisational structure, social media details, web links, and LinkingYou links
- Student clubs and societies
- Name, purpose, web links, and social media details
If you’d like to be involved in our requirements gathering — as either a data provider or consumer (or both!) — please get in touch by emailing email@example.com.