Datasmiths and Data Forging: Ontology, Datafied Living, and Cyberinfrastructures

• In Event: Datafied Living: The Everyday of Datafication

Abstract

The notion of a ‘cyberinfrastructure’ goes back to 1980-90s research on high computational capacity supercomputers and common goal distributed grid computing. Today, cyberinfrastructure refers to largescale computing networks used by interdisciplinary researchers and administrators to gain novel insights via the high-speed sharing of data resources. As one can tell by the name, cyberinfrastructure is a combination of the words ‘cyber’ (relating to computation) and ‘infrastructure’ (basic organizational structure)—in simple terms, ‘cyberinfrastructure’ means something like the ‘organization of computing resources.’ An update on older, primarily physical infrastructures (highways systems, bridges, rail lines, the electric grid, etc.), cyberinfrastructures are a prime example of what Couldry and Hepp refer to as ‘the mediated construction of reality’ (2016) through virtual and physical data connections between supercomputers, databases, software, researchers, and institutions via the internet. The alleged end goal of most cyberinfrastructures is to spur innovation, scientific discovery, and knowledge growth by connecting massive amounts of distributed data across great distances, allowing researchers and administrators to analyze each other’s data and conduct new experiments or test theories. Building on the work of critical cyberinfrastructure researchers (Ribes and Lee 2010), this paper focuses on the issue of semantic data interoperability (Ribes and Bowker 2009) and the potentially harmful political implications of cross-sector data harmonization. Every day, disparate data relating to social entities and relations in areas such as health, governance, intelligence, and economics proceed through cyberinfrastructures via highly standardized metadata annotations known as ontologies (Arp, Smith and Spear 2015). Ontologies sometimes sort social data by creating normative categories, into which heterogeneous data must fit, a practice that can sometimes lead to biases and errors in big data organizing and reasoning. In this paper, I analyze several ontology projects that process large amounts of data relating to social entities and relations across different community databases. Throughout, I refer to the practitioners involved as ‘datasmiths’ who are involved in the practice of ‘data forging’ with the aim of streamlining messy, socially determined data categories. The goal will be to provide a critical assessment of ontologies, ontologists, and ontology-making practices.

Author

• Andrew Iliadis, Temple U - Media Studies & Production