by Kim Martin, on behalf of the HuViz team
Although I’ve been aware of the Pelagios Project for a few years now, I only had my first in-depth encounter with Recogito during a workshop at DH2018 in Mexico City. I was impressed by the slick interface of the tagging tool, the various file types available for output and the visual aspects of creating relationships between the tags (a feature that was released just prior to the workshop I attended). Though I understand the benefits of displaying the gazetteer data on a map, my own expertise lies outside of the geospatial realm, and I was left wondering what else users might want to do with the textual material they are tagging in Recogito. Having spent the last two years as the project manager for HuViz (a digital tool for visualizing semantic relationships and ontologies) and a member of the Canadian Writing Research Collaboratory (CWRC) ontology team, I am most familiar with the RDF/Turtle data format, and decided to test what this might look like in HuViz. I uploaded a single TEI-encoded file from the Records of Early English Drama (REED) London to Recogito, tagged a few other people and places within this text using Recogito, exported the .ttl file, and uploaded this to HuViz. The results are shown here:
The pink nodes at the top left of the circular shelf are literals resulting from previous TEI tagging (performed in Oxygen), the light blue nodes are annotations made to the text using Recogito, and the light green are the tags themselves. All of the annotations are strings, making the graph difficult for humans to interpret. I left the 3-hour workshop intrigued by Recogito, but unsure how I would put the resulting datasets to use.
In December of 2018 I attended Linked Pasts IV with Susan Brown (Director, CWRC) and found that more and more people were using Recogito to tag people, places, and the relationships between these entities. It became evident to me that others would benefit from having other ways, beyond a map, of visualizing annotations they created in Recogito.
We brought Rainer and other members of the Pelagios team into the conversation, and concluded that a bridge between the two projects would be fruitful. The catch: from Rainer’s point of view, it would be best if HuViz could read JSON-LD files (at the moment it reads only RDF/Turtle). So I returned from Linked Pasts with some queries for our developer, Shawn Murphy of Nooron Collaboratory, whose positive feedback resulted in an application for a Pelagios Small Grant, which we are thrilled to have received for 2019!
The idea of this mini-grant is to ensure compatibility between two digital humanities projects: Pelagios and CWRC. Focussed linked-data-oriented tools have come out of both projects: Pelagios developed Recogito for tagging documents with place names and visualization of that data on maps, while CWRC has developed the CWRC-Writer TEI and RDF editor focused on annotating texts with named entities more generally, and the HuViz Linked Data Explorer for visualizing relationships between entities. Through Interchange, we are working together to connect siloed datasets by making the tools more interoperable, benefitting from each project team’s specific knowledge, and each tool’s specific affordances.
Over the next 9 months, you’ll see posts from the HuViz team highlighting the steps we are taking towards our final goals:
- Making HuViz read JSON-LD files
- Offering a human-friendly view of Web Annotation data in HuViz.
- Creating modular ‘cookbooks’ for the interchange process, which consists of:
* transferring XML documents from Recogito to CWRC-Writer,
* pushing data from Recogito to CWRC-Writer,
* transferring data from CWRC-Writer to HuViz; and
* taking data from CWRC-Writer back into Recogito.
We look forward to hearing feedback from users as we progress in these tasks. If you’d like to participate in user-testing throughout this process, please let us know.
