Pelagios’ Linked Data Workshop

Last week the Pelagios team held a workshop on linking open data and ontologies. The event started with introducing the various team members. Being a wide-spread and international team, this was also the first time several of the members had met each other. The talks, relating to linked data for referencing ancient places, brought people from various backgrounds so it was helpful and also insightful when speakers said where they were coming from as they started their talk. Geographers, classicists, developers, and general researchers or data-lovers, someone could make quite a venn-diagram of us all. It was all recorded, and concluded that the presentation slides should be shared, I’ll update this post with a link when I hear they’ve been uploaded.

The first section of talks, “Issues”, had started with the problems encountered when we want to refer to the same location in our documents or systems. Athens was a regularly used example as it could refer to various different cities around the world. Time is also a field worth referring to, especially in connection with places that move, disappear, or change names over time.

For those not familiar with the concept of Linked Data, Jeni Tennison from data.gov.uk gave a good introduction and rules on what Linked Data should be. I recommend watching the recording, even if you’ve been working with linked data for a while, it might make you reconsider if you’re linking is strict enough and your data purposeful.

After  lunch, ontologies were discussed. Thinking about how to organise your data of terms, and seeing how others have done it. Here the subject of time continued to flow through the talks and how it’s affected previous projects. John Goodwin, from the Ordnance Survey, told us how they are providing data to be linked to so a common reference can be used for modern locations.

In the final section, titled Methods, we heard how other projects had been working on linking data together. Claire Grover talked about the Edinburgh Geoparser that scanned texts for locations and used various gazetteers to translate those into geospatial co-ordinates. This followed on to talk about the jiscEXPO funded Chalice, which is looking at historic documents from the Historic Place-Name Society so that translations can be made linking between modern names and old names for places.

The issue of fuzziness was bought up as Ceri Binding talked about working with archaeological data and time references. The Stella application shows how time periods (ages, eras, reigns, etc) relate to each other (before, after, during, at the end,…) by giving each period a minimum start year and maximum end year. Unlike monarchy reigns, some ages and time periods we can’t set the exact years to define them, further complicated by ages emerging across the world at different times at different rates. This all makes it tricky to create relations, as you make it hard to say if one time period led onto another period, or if it ended some years before the other started.

In the last talk, Eetu Mäkelä add a further level of questions by  asking if you say a building was built 2000-2010, does that mean it took 10 years to build or is it only known it was built sometime in that decade.

The day was insightful and mentioned multiple projects and services that could be good for Pelagios to use or learn from. The time is the obvious subject that could be picked up on. Both time and location are closely linked with each other but through their abstract concepts a lot can be learnt from referencing time and put into place for building up place name references.

Update: The Pelagios team now have a page about the workshop with photos. the recording will be added to that page when available.