Will Historians Ever Have Big Data?
Digital history has spawned many great individual projects, and proven its value as both a methodology for the interrogation of sources and as a medium for the presentation and communication of research results. But the experiences of projects building re
- PDF / 168,115 Bytes
- 15 Pages / 439.37 x 666.14 pts Page_size
- 27 Downloads / 207 Views
)
Trinity College Dublin, Dublin, Ireland [email protected]
Abstract. Digital history has spawned many great individual projects, and proven its value as both a methodology for the interrogation of sources and as a medium for the presentation and communication of research results. But the experiences of projects building research infrastructure for historical research raise the question of whether these methods can scale toward the realisation of historical ‘big data,’ or whether there are hindrances to this goal inherent in our current conceptualisation of the intersection between historical methods and computational ones. This paper discusses a number of the current barriers discov‐ ered by large-scale historical research infrastructure projects, including hetero‐ geneous conceptions of what data is, hidden elements of data and the epistemics of humanities research. At the project level, these issues can be managed, but if digital history is to scale and grow to fit the infrastructural capability available to it, then a revisiting of some of the conceptual underpinnings of digital historical studies will be required. Keywords: Cliometrics · Epistemics · Cultural computing · Digital humanities · Big data · Provenance · Authority
1
Introduction “Less Guessing. More Knowing. Analytics, Delivered.” Accenture Advertisement, Dublin Airport, April 2012 “Analysing Big Data, That’s the secret to living happily ever after.” Winton Global Investment Management Advertisement, London Underground, May 2015
As a society and as a research community, we seem to worship ‘big’ data. But like any other product of the human race, our datasets are socially and individually constructed, and prone to error and bias - indeed in many cases it is this very individuality of datasets that is the mark of their provenance, the knowledge organisation framework their creator applied, and an inherent part of their utility as a foundation for knowledge creation. Even when they are ‘clean’ to a ‘gold standard’ or ‘open’ to a ‘five star’ rating, datasets of any complexity, analogue or digital, remain objects that need to be viewed in the context of their creation. © IFIP International Federation for Information Processing 2016 Published by Springer International Publishing AG 2016. All Rights Reserved B. Bozic et al. (Eds.): CHDDH 2016, IFIP AICT 482, pp. 91–105, 2016. DOI: 10.1007/978-3-319-46224-0_9
92
J. Edmond
So long as the complexities of data and their sources remains visible to the user, this ‘human all too human’ variation maintains its capacity to be a strength, rather than a weakness. But statistical and engineering imperatives have fostered approaches based on the assumption that the increasing scale of data is a hallmark of increased knowledge, of authority, perhaps even of a sort of ‘truth.’ To further increase the scale of integration of data into truly big data is to hide that complexity, cultural specificity and the social constructedness of data in a ‘black box,’ and to flatten the nuances inherent in it that may be essential to it
Data Loading...
