2020-09-21, 17:10–17:40, Room 1
This talk aims to provide an historical context for the renewed interest in Discrete Global Grids (DGGs) by exploring what they do well and what they don't. In particular, we'll take a Python-centric look at what happens when we need to assimilate gridded data with data reported at the level of traditional administrative boundaries.
Discrete Global Grids (DGGs) are an increasingly popular tool for collecting, indexing, and analyzing geospatial data. They offer researchers and professionals an enormous amount of consistency, flexibility, and control over the data aggregation process while avoiding many of the biases imposed by more traditional geographic boundaries (e.g. the Modifiable Areal Unit Problem (MAUP)).
Once one moves beyond the realm of aggregating point data, however, the benefits of leveraging these globe-spanning mosaics become more tenuous. In particular, the process of assimilating grid-based data with data bound by geometrically irregular shapes (e.g. administrative boundaries) often undermines the benefits offered by DGGs while at the same time creating new, potentially avoidable challenges. And because most sociodemographic data are reported at the level of administrative boundaries, anyone wishing to derive meaningful insights from gridded data sources will be forced to make certain assumptions when they eventually confront these incompatibilities.
This talk will situate the modern DGG in an historical context, and use specific examples from the GeoPython ecosystem to contemplate its return to fashion. We will explore the promises and pitfalls of using a discrete global spatial grid to perform geospatial analysis, and see what Python-based tools and techniques are available to researchers who find themselves having to bridge the gap between gridded and irregular geographic boundaries.