Encoding Collective Knowledge, Instructing Data Reusers: The Collaborative Fixation of a Digital Scientific Data Set

Abstract

This article provides a novel perspective on the use and reuse of scientific data by providing a chronological ethnographic account and analysis of how a team of researchers prepared an astronomical catalogue (a table of measured properties of galaxies) for public release. Whereas much existing work on data reuse has focused on information about data (such as metadata), whose form or lack has been described as a hurdle for reusing data successfully, I describe how data makers tried to instruct users through the processed data themselves. The fixation of this catalogue was a negotiation, resulting in what was acceptable to team members and coherent with the diverse data uses pertinent to their completed work. It was through preparing their catalogue as an ‘instructing data object’ that this team seeked to encode its members’ knowledge of how the data were processed and to make it consequential for users by devising methodical ways to structure anticipated uses. These methods included introducing redundancies that would help users to self-correct mistaken uses, selectively deleting data, and deflecting accountability through making notational choices. They dwell on an understanding of knowledge not as exclusively propositional (such as the belief in propositions), but as embedded in witnessable activities and the products of these activities. I discuss the implications of this account for philosophical notions of collective knowledge and for theorizing coordinative artifacts in CSCW. Eventually, I identify a tension between ‘using algorithms’ and ‘doing science’ in preparing data sets and show how it was resolved in this case.

GeoDeepShovel: A platform for building scientific database from geoscience literature withAIassistance

“…All the literature uploaded into GeoDeepShovel is all automatically parsed with Grobid (GRO, 2008-2021 and F I G U R E 8 GeoDeepShovel System overview. GeoDeepShovel consists a PDF prasing module, a backend server including some artificial intelligence models and a interactive graphical user interface.…”

“…For each uploaded document, GeoDeepShovel uses multiple parsing tools (e.g. Grobid (GRO, 2008-2021, Science Parse and PdfFigures 2.0 (Clark & Divvala, 2016b)) to independently extract its meta-information and mix all the information with a voting mechanism. The meta-information of papers (e.g.…”

“…Geoscience research often relies on a large amount of exploration data. The shifting of the data‐driven research paradigm raises new requirements for researchers to build scientific databases (Hoeppe, 2021) in Geoscience (Bergen et al, 2019). Scientific database is a collection of structured and verified research results that consist of various numeric, word‐oriented or image‐organized data, which plays a central role in data‐driven research (National Research Council, Division on Engineering and Physical Sciences, Commission on Physical Sciences, Mathematics, and Applications, Committee for a Study on Promoting Access to Scientific and Technical Data for the Public Interest et al, 2000).…”

DeepShovel: An Online Collaborative Platform for Data Extraction in Geoscience Literature with AI Assistance

“…Some researchers noticed and analyzed the problems in team collaboration that existing platforms are facing. Hoeppe [22] mentions that the large scientific databases often require large teams to extract and clean the data, and it is a complex task that needs to be done through CSCW. Schmidt [45] demonstrates that task distribution, allocation, and interrelating of 'distributed individual activities' are some important issues in team collaboration.…”

KnowledgeShovel: An AI-in-the-Loop Document Annotation System for Scientific Knowledge Base Construction

“…Scientific research is currently advancing faster than ever before, with tens of thousands of scientific studies published every day in various forms such as papers, preprints, and datasets. Scientific knowledge bases [16,23], a collection of structured and verified research results that consists of various numeric, word-oriented, or image-organized data, emerge in this context and bring entirely new approaches and opportunities to scientific research. Researchers in many disciplines uses AI techniques and the scientific knowledge bases, often constructed from the published literature, to drive scientific discoveries [38,45,46], such as Geoscience [10,64], Medicine [9], Biology [3], Chemistry [50].…”