Hacking Web Maps-T
The Web Maps-T working group aims to enhance the ability to visualise geospatial and temporal Linked Open Data on web maps. The group is coordinated by Gethin Rees and Adi Keinan-Schoonbaert, see this previous post for more details.
The white paper written by the group is now available.
As a first step, we held a hack workshop in September at the British Library to scope applications to visualise and work with the GeoJSON-T standard. Participants at the workshop included from the Rainer Simon from the Austrian Institute of Technology, Karl Grossner from University Of Pittsburgh, Neil Jakeman from King’s College Digital Lab, Alex Butterworth and Simon Wibberley from Sussex University’s Humanities Lab alongside Mia Ridge and Olivia Vane from Digital Scholarship at the British Library. They came to the workshop with a variety of datasets derived from the art, archaeology and travelogues from the classical world for example, as well as the collections of the British Library, including contributions from the Endangered Archives Programme, Georeferencer and Two Centuries of Indian Print.
Throughout the event we collaborated using repositories in the Pelagios GitHub organisation and Google Docs. After introductory talks, first by Gethin Rees on user stories and use cases for Web Maps-T, and second by Karl Grossner on GeoJSON-T and the Linked Traces app, participants introduced their own projects and use cases for geospatial and temporal visualisation.
Participants then coalesced around several tasks and divided into groups to work. These tasks included:
· MoSCoW assessment of Web Maps-T app. This prioritisation technique divided potential features of the app into Must, Should, Could and Won’t. For example, our app must have a timeline, a map, and work with GeoJSON-T. On the other hand it could use animation for visualisation, and won’t include a text box querying for plain English.
· Classification of datasets and visualisation types. We wanted to explore several forms of visualisation and it quickly became clear that some were better suited to certain types of datasets than others. For histogram, linear journey, period horizontal bars, valid time intervals and beeswarm visualisations we recorded data type, size, examples and pros and cons.
These activities were documented in detail and will form the basis of the white paper. Participants that wrote code worked on:
· WhenJSON — a front-end JavaScript utility library for manipulating temporal data in the GeoJSON-T format (the ‘when’ object). The utility could, for example, help in calculating the interval that separates two data sources.
· A Minimum Viable Product (MVP) web map with a time-slider component to test visualisation types for different dataset classifications. This work borrows heavily from the work of Jonathan Skeate. The MVP allowed us to see how datasets looked when visualised using different methods such as histogram or time bars.
· Scoping the adaption Karl Grossner’s Linked Traces app to be used with any dataset and to add a time slider. This could be a more robust and presentable solution than our adaptation of Skeate’s Leaflet Timeline.
Next steps for the group are to write a white paper and to attend the Linked Pasts conference in Bordeaux. All the workshop participants had projects where they were keen to apply Web Maps-T and there must be many more use-cases out there. However, the path from hack event to an open-source, production-ready component that can visualise any GeoJSON-T data we throw at it is not straightforward. Like any open-source project, success rests on people and projects requiring the component and standard for their day-to-day work. There are plenty of potential applications in the Linked Pasts community and beyond, the main resources that we now require are developer time and GeoJSON-T datasets. We will continue to encourage others to contribute, perhaps you have potential applications. Get in touch at visualisation[at]pelagios.org
