Two-hour technical virtual workshop on Multiscale Geographically Weighted Regression at University Consortium for Geographic Information Science workshop series

Abstract

This half-day workshop will introduce participants to the local spatial modeling technique MGWR. The presenter will introduce the fundamental theoretical and technical foundation required to understand the tool and then walk through an application of the method. Participants will be given a hands-on introduction to the types of research problems that can be investigated with the tool. Participants will also learn how to prepare data for implementation in the software followed by critical visualization of the results. The workshop will consist of three parts. In the first part of the workshop, the presenter will introduce the overarching conceptual research problem being addressed through the development of the technique and will subsequently introduce the technical methodology. Building on this foundation, participants will then be guided through a worked example implementing a hedonic housing price model for Seattle, WA. The presenter will guide participants through hypothesis formation, development of the spatial model, and implementation of that model in the open-source MGWR software. The presenter will then introduce the various outputs obtained from fitting the model and will explain the interpretation of the results. In the final part of the workshop, the presenter will demonstrate a visualization protocol developed as a Python based Jupyter Notebook that will allow participant to graphically display results. The presenter will conclude the workshop with a short discussion on other examples applying the technique and future research avenues, and invite participants to work with their own datasets. The MGWR software is free to download (MGWR software) and workshop datasets will be made available to participants prior to the workshop. In addition, the presenter will make available all Jupyter Notebooks used during the workshop (e.g. the visualization protocol) to facilitate future participant application/exploration with the software.

Mehak Sachdeva

Urban Science Faculty Fellow

My research interests include spatial statistical methods with applications in the social, urban and environmental sciences.