Book Image

Hands-On Geospatial Analysis with R and QGIS

By : Shammunul Islam, Brad Hamson
Book Image

Hands-On Geospatial Analysis with R and QGIS

By: Shammunul Islam, Brad Hamson

Overview of this book

Managing spatial data has always been challenging and it's getting more complex as the size of data increases. Spatial data is actually big data and you need different tools and techniques to work your way around to model and create different workflows. R and QGIS have powerful features that can make this job easier. This book is your companion for applying machine learning algorithms on GIS and remote sensing data. You’ll start by gaining an understanding of the nature of spatial data and installing R and QGIS. Then, you’ll learn how to use different R packages to import, export, and visualize data, before doing the same in QGIS. Screenshots are included to ease your understanding. Moving on, you’ll learn about different aspects of managing and analyzing spatial data, before diving into advanced topics. You’ll create powerful data visualizations using ggplot2, ggmap, raster, and other packages of R. You’ll learn how to use QGIS 3.2.2 to visualize and manage (create, edit, and format) spatial data. Different types of spatial analysis are also covered using R. Finally, you’ll work with landslide data from Bangladesh to create a landslide susceptibility map using different machine learning algorithms. By reading this book, you’ll transition from being a beginner to an intermediate user of GIS and remote sensing data in no time.
Table of Contents (12 chapters)
GRASS, Graphical Modelers, and Web Mapping

Landslide susceptibility modeling

For landslide susceptibility analysis, we will follow these steps:

  1. We will load a digital elevation model (DEM) from SRTM (provided in the Data folder of Chapter10).
  2. Compute the slope from this DEM.
  3. Reproject each file to the same projection system so that we can compare one with another.
  4. Extract the elevation and slope values corresponding to the landslide location using the DEM file and the computed slope.
  5. Classify areas as safe or unsafe using the range of slopes within which landslides have occurred.
  6. Get random points in the safe zone and extract their elevation and slope values.
  7. Using the elevation and slope values from both the unsafe and safe locations, fit a logistic regression model.
  8. Using the coefficient of the fitted model, make a landslide susceptibility map.