A course on Geographic Data Science

Data Science (Donoho, 2017) has become one of the most demanded skills thanks to an explosion in the availability of data (Kitchin, 2014). Most of these new sources are, directly or indirecly, geographic in that they can be related to a particular location on a map. However, the vast majority of data science resources available currently ignore the spatial dimension of data, particularly when it comes to the more analytic set of methods covered. At the same time, traditional resources for teaching the handling, visualisation, and analysis of geographic data are based on a paradigm that emphasises graphical interfaces and “point-and-click” software packages. This approach, although valid, limits the flexiblity with which the analyst can effectively move from data to insights, and is more difficult to connect with and benefit from modern advances in both data tools and workflows. This paper presents a pedagogical bridge between the “spatially unaware” set of practices emerging from Data Science, and more traditional resources designed to teach spatial analysis within a Geographic Information Systems (GIS) environment.