EcoCor: The New Generation
What is the project about?
Our “EcoCor: The Next Generation” Challenge consists of several subchallenges:
- “EcoCor100”: We will significantly grow our (German) corpus of ecocritically relevant literary (prose) texts, so that it includes at least 100 texts in it
- “Dance of the Taggers”: We will examine and compare different approaches and tools for entity annotation (focusing on animals and plants).
- “EcoCor Infrastructure”: We want to redesign the EcoCor infrastructure so that it can accommodate different annotations and output them via the API. We also urgently need a new visualization approach for the word clouds.
- “One Big Beautiful Network”: Using the EcoCor API, we want to generate network data, including for individual narrative texts, as well as a merged network for all literary narrative texts in the German-language EcoCor. We then want to visualize this.
References and links
What is your research question?
Which literary texts are ecocritically relevant from an either qualitative or quantitative perspective and how can we process and present them via EcoCor in an accessible way, which adds value to research and, more broadly, to the understanding of literary texts.
How is the project and/or case situated?
The project combines (Computational) Literary Studies, Computational Linguistics, and Informatics. It is a corpus compilation project with two main perspectives. On the one side, we want to gather at least 100 texts that are of interest for (digital) Environmental Humanities.
We use second-hand criticism as well as frequency analysis to find texts that have already been part of ecocritical studies or could be of interest for future work. We follow ideas from data feminism and aim at widening the canon towards gender parity between male and female authors.
From a technical perspective, we develop a pipeline to add raw texts to the EcoCor platform and enhance it with (multiple) annotation layers. We thus provide raw texts as well as a first annotation layer created using a dictionary lookup to highlight animals and plants and a second one based on Machine Learning to annotate animals, plants, and habitats.
In sum, EcoCor100 is a Digital Humanities project centering corpus creation and publication.
What methods, data sets, and tools are used?
Distant Reading, Network Analysis, Machine Learning, LLMs
Who is part of the team?
Mark Schwindt (Ruhr-University Bochum)
Daniil Skorinkin (University of Potsdam)
Sören Barkey (University of Potsdam)
Henny Sluyter-Gäthje (University of Potsdam)
Carsten Milling (University of Potsdam)
Peer Trilcke (University of Potsdam)
Ingo Börner (University of Potsdam)
Thomas Nikolaus Haider (University of Passau)
Mareike Schumacher (University of Regensburg)
Clara Helmig (University of Regensburg)
Corinna Käb (Eberhard Karls University Tübingen)
Bianca Ottenberg (University of Trier)
Rebecca Daniel (University of Trier)
Renée Ridgway (Aarhus University)
Clara Runa Schlör (FU Hagen)
Frank Fischer (Freie Universität Berlin)