Spatial R: A Quick(ish) Introduction

Author: Colin Chisholm

Initiated: Oct. 27, 2018

Last Update: Oct. 28, 2018

This was my first examination of looking at spatial data in R. Since the creation of this information I have found many other resources.

One of the best resources I have found for R coding and spatial data is Geo-computation with R by Robin Lovelace, Jakub Nowosad, and Jannes Muenchow.

Purpose

To create a series of R scripts starting with basic spatial data imports and analysis. This project is not intended to create original or new material for the R community but to be used for self training and to be provided as a resource for others. Additionally, my goal is not (at least to start) worry about cartography in R but to focus on data extraction and manipulation.

Introducing R

The purpose of this project is not to introduce people to the basic functionality of R. Readers are encouraged to visit sites like such as

• Data Camp Intro R provides a free intro to R course
• Data Camp -- Cleaning Data in R is a free courses or at least the first modules are free. This course focuses on dataframe manipulations, sorting and organizing your data.
• R-Bloggers covers numerous topics on R. The posts range from the simple to complex issues and usually include detailed descriptions of how to do things. There are also many links on this site
• Google Search: "R packagename"

The only two tips I will suggest here are: First and most importantly, to add lots and lots of #comments_ as these will assist you in the learning process and help you reconstruct what you were doing after you leave some of your coding for a period of time. Second, there is no shame is taking snippets of other peoples code on the web ... just comment where you found it including a description of what it is used for.

Introduction to Spatial Data in R

To some extent I am reinventing the wheel here. There are other sources to find this information. However, they can at time be too thorough. Here I am expecting that readers have a decent understanding of GIS data types (e.g. from ESRI Arcmap, Global Mapper, of QGIS experience) and simply want to start using some of the tools, repeatable script processing that R provides. The National Ecological Observatory Network (NEON) provides some excellent and thorough tutorials:

• NEON's Introduction to Vector Data in R
• NEON's Raster Data in R -- the Basics

---

Project Outline / Map

This project will include the sections listed below. Starting with basic vector and raster import, display, and data manipulation. These will be followed by integration of raster and vector data sets. Currently, Random Forest Modeling is planned as this is a major component of ongoing research projects I am involved with (PEM Project).

• [00] Libraries
• [10] Vector data
• [20] Raster data
• [30] Integration of Raster and Vector data
• [50] Random Forest Modeling
• [90] Code Snippets and My Functions

---

Essential Commands -- The Short Version

Import/Export

Vectors

Importing spatial data readOGR. For importing non spatial data myData.csv which assuming spatial data exists a dataframe can be converted to a Spatial data type via SpatialPointsDataFrame(myData[,x:y], mydata, proj4string = myCRS)

Data can be exported as a shapefile using writeOGR(myData, PATH, NAME, driver = "ESRI Shapefile"). Note that other data types are possible see: https://www.gdal.org/ogr_formats.html.

Data Examination

Unique to spatial data: crs() extent(). Otherwise use common R dataframe commands such as summary() str() length() and many more. Essentially consider the spatial vector data as just a data frame with spatial information attached.

Rasters

Import a raster raster(FileName)

Manipulations

Change or assign CRS spTransform(Layer, TargetCRS) Crop raster data crop(rLayer, Extent)

Queries

df <- extract(Raster, AOI) will place all of a raster's pixel values into a dataframe. When the AOI is a polygon is will extract all the overlapping values.

Query raster data using points - a quick process/example

#Import the data
myPoints <- readOGR(PATH, FILENAME)      #import vector data (e.g. points)
myRaster <- raster(FILENAME)             #import raster data (filename including path)

# extract the mean of a  11.28m plot around the point (circular)
myPoints$RasterInfoMean <- extract(myRaster, myPoints, fun = "mean", buffer = 11.28)
# OR extract exact the exact point values
myPoints$RasterInfoPoint <- extract(myRaster, myPoints)

#export the data
writeOGR(myPoints, PATH, FILENAME, driver = "ESRI Shapefile")

---

00. Libraries

For those new to R or scripting in general it is important to realize that the base scripting/programing language is expanded using additional libraries. Note other languages call these packages (e.g. Python). In the files/scripts provided in here the additional libraries are listed at the top of the file with a short description. The intent will be to list all of the libraries used throughout this entire package set. However, unless they are explicitly needed in the current script they will be commented out.

---

Vector Layers

10. Vector Data - Working with Single Layers

This script walks through how to import vector data, explore the data types and attributes, and plot the data based on the attributes. Examples include Line, Polygon, and Point datatypes.

11. Vector Data - Working with Multiple Layers

This script, again following the NEON tutorials, demonstrates how to work with multiple layers. Adding additional layers to a plot is a simple process of using the additional flag of add = TRUE in a plot() call. The greater tricks are related to how to build legends.

12. Vector Data - Projections and Transformations

Transformation of data from one projection to another is quite straight forward using spTransform(). For example:

ProjectCRS <- crs(States)                #sets CRS value of Layer States
Tower <- spTransform(Tower, ProjectCRS)  #Transforms Tower Layer to States the ProjectCRS

13. Import CSV Vector Data and save to Shapefile

Process:

1. Reading in the data using read.csv examine the data for geographic information `str(data)``
2. Look up the CRS using the web. For example: http://www.spatialreference.org/
3. OR use spTransform if you have a layer with the correct projection in it (easiest)
4. Convert the CSV from a table to spatial object
5. Export as shapefile

#Convert loaded dataframe (containing spatial data) to a spatial object
dataPoints <- SpatialPointsDataFrame(dataPoints[,1:2],    #The x y coordinates
                                 dataPoints,              #The Data to assign
                                 proj4string = ProjectCRS #Assign the CRS
                                 )
#Export as shapefile
writeOGR(dataPoints, "E:/workspace/SpatialR/exports", "PlotData", driver = "ESRI Shapefile")

The NEON lab also discusses extent and changing the extent of a plot as per this example

# adjust the plot extent using x and ylim
plot(myData,
     main="Title",
     xlim=c(xmin,xmax),
     ylim=c(ymin,ymax))

---

Raster Data

21. Import and crop raster data

This module examines how to import a raster which is a simple call raster(FileNameInclPAth). It then examines how to clip the raster layer to an area-of-interest. This can be important as rasters can take large memory requirements. To determine which vector layer to clip to an simple function is included to calculate the extent area of a layer. To crop a layer crop(Layer, Extent) note: that the layer can be any spatial object and the Extent can be another layer or a specified extent

Raster / Vector Integration

31. Extract Raster information

This script loads several layers (polygon, line, point, and raster), clips the raster to an extent, and extracts the data as a whole or by vector point overlays. The key new function is extract().

90. Code Snippets

This is a file where I am storing various code snippets and functions that can be copied to future projects. Functions include:

• ExtentArea(Layer) which returns the area of the extent of the layer in question.