Management of Biotic Interaction Data in the Encyclopedia of Life

Poster was presented at the (virtual) 4th Annual Digital Data in Biodiversity Research Conference at Indiana University 1-3 June 2020 . Abstract: Species interactions (predator-prey, parasite-host, pollinator-plant, etc.) are an important component of TraitBank, the structured data resource developed by the Encyclopedia of Life (EOL, eol.org/traitbank). TraitBank gains access to these data through Global Biotic Interactions (GloBI, globalbioticinteractions.org), an open-source infrastructure for the efficient integration & sharing of interaction data from diverse sources, including scientific literature/databases, specimens from natural history collections, text mining & citizen science projects. Interaction data are featured on EOL taxon pages in data tables & food web visualizations, and human- & machine-accessible interfaces are provided to query & download data. To facilitate harvesting by EOL, GloBI supplies data summarized by reference in an association extension based on the Darwin Core standard. The data are further processed in a custom EOL connector that addresses taxonomic redundancies & applies a series of validation rules designed to catch data errors, for example, a claim of a plant pollinating an animal. To support efficient feedback about potential data problems, validation failures result in the creation of new refuted interaction records that challenge the original claim. This EOL curated dataset of refuted interaction records is then indexed by GloBI, so other GloBI users have the option to identify data records that are incorrect or controversial. Included audio files contain an introduction and commentary on the poster by Katja Schulz. See also https://www.idigbio.org/wiki/index.php/4th_Annual_Digital_Data_Conference,_Indiana_University