Spatial Modelling for Data Scientists

A course created by Francisco Rowe, Dani Arribas-Bel.

The course covers

how to analyse and model different types of spatial data as well as gaining an understanding of the various challenges arising from manipulating such data.

Note that this resource only provides the practical coding portion of the course. The theoretical lessons are not available.

The activities can be completed in 10-12 laboratory sessions of 1-2 hours depending on the competence of the participants.

Stated aims

• build upon the more general research training delivered via companion modules on Data Collection and Data Analysis, both of which have an aspatial focus;
• highlight a number of key social issues that have a spatial dimension;
• explain the specific challenges faced when attempting to analyse spatial data;
• introduce a range of analytical techniques and approaches suitable for the analysis of spatial data; and,
• enhance practical skills in using R software packages to implement a wide range of spatial analytical tools.

Learning Outcomes

• identify some key sources of spatial data and resources of spatial analysis and modelling tools;
• explain the advantages of taking spatial structure into account when analysing spatial data;
• apply a range of computer-based techniques for the analysis of spatial data, including mapping, correlation, kernel density estimation, regression, multi-level models, geographically-weighted regression, spatial interaction models and spatial econometrics;
• apply appropriate analytical strategies to tackle the key methodological challenges facing spatial analysis – spatial autocorrelation, heterogeneity, and ecological fallacy; and,
• select appropriate analytical tools for analysing specific spatial data sets to address emerging social issues facing the society.