Algorithmic Geographies: Big Data, Algorithmic Uncertainty, and the Production of Geographic Knowledge

Drawing on examples from human mobility research, I argue in this article that the advent of big data has significantly increased the role of algorithms in mediating the geographic knowledge production process. This increased centrality of algorithmic mediation introduces much more uncertainty to the geographic knowledge generated when compared to traditional modes of geographic inquiry. This article reflects on important changes in the geographic knowledge production process associated with the shift from using traditional “small data” to using big data and explores how computerized algorithms could considerably influence research results. I call into question the much touted notion of data-driven geography, which ignores the potentially significant influence of algorithms on research results, and the fact that knowledge about the world generated with big data might be more an artifact of the algorithms used than the data itself. As the production of geographic knowledge is now far more dependent on computerized algorithms than before, this article asserts that it is more appropriate to refer to this new kind of geographic inquiry as algorithm-driven geographies (or algorithmic geographies) rather than data-driven geography. The notion of algorithmic geographies also foregrounds the need to pay attention to the effects of algorithms on the content, reliability, and social implications of the geographic knowledge these algorithms help generate. The article highlights the need for geographers to remain attentive to the omissions, exclusions, and marginalizing power of big data. It stresses the importance of practicing critical reflexivity with respect to both the knowledge production process and the data and algorithms used in the process.