Data and Purpose : The relationship between people, place and data

Hollinshead, Jan and Harper, Richard (2025) Data and Purpose : The relationship between people, place and data. PhD thesis, Lancaster University.

Text (2025HollinsheadPhD)
DataAndPurpose_Hollinshead.pdf - Published Version
Available under License None.
Download (10MB)

Abstract

This thesis explores the practices of data science from an epistemological perspective. It starts with a look at the concept of data and critiques the claim that data “speak for themselves”. Rather, it suggests, people embed meaning into data through the purposes for which they are created, shaped, and used. One such purpose of data is in representing place, and the wider cross-disciplinary concept of place is discussed. Place is not just described by data but also provides context for data, and should, therefore, just as with data, be defined by the purposes for which it is used. If data practices can be thought of as the combination of data, place and purpose, then this should be observable in everyday life. Empirical examples are offered of individuals whose data practices use personal informatics data and technology in artful and reflexive ways to construct meaning about themselves and their places. Assuming individuals’ data practices to be similar to organisational sense-making, this leads to an exploration of the relationship between the role of data and data “products” by organisations. The way such data are generated needs to be thought of as part of data practices, and the hidden work and the artfulness involved in generating data in real-world situations, are made visible through an auto-ethnographic case study of data practices. The thesis concludes that data, purpose and place are entangled in the data practices of individuals and organisations alike, and they build on everyday sense-making practices to give ways of knowing. That this is not included as part of the data science epistemology and training perpetuates the data discourse and the mis-representation of data. The implications that this has on the indexicality and nature of data science within academia, and for data-driven AI technologies in general is commented on.