Goal: Make our first map!

 

Steps for lab:

  1. Import a shapefile into QGIS/R (2020 election results in MI)
  2. Color the electoral precincts
    blue and red by vote share
  3. Save the map as an image
  4. Create and map a new variable: vote margin

 

Problem set:

  • Repeat these steps for election data from Massachusetts (in QGIS or R)


We’ll make this

Data

 

For this exercise we will be using a precinct-level dataset on 2020 presidential election results for the state of Michigan, mi_2020.

 

Let’s get started:

  • download Lab01PS01.zip from Canvas (Files \(\to\) Labs \(\to\) Lab01PS01.zip)
  • extract zip file to a folder on your hard drive
  • the file we need is this: Lab01PS01/Data/mi_2020.geojson

 

Dataset citation info:

Voting and Election Science Team, 2020, “2020 Precinct-Level Election Results”, Harvard Dataverse, V41. doi.org/10.7910/DVN/K7760H

QGIS

Let’s open QGIS!

 

Get ready to rock!

Pro tip: Save your progress!
Go to Project \(\to\) Save As...

QGIS can save the state of your workspace into a project file

  • the extension for QGIS projects is .qgz, a compressed file format
  • the .qgz file contains information on your project’s layers, symbolization and styles, projections and print layouts
  • it does not contain your data files, or any temporary files you created
  • this can be very helpful if QGIS crashes, so you can pick up where you left off

Introduction and Demo

  1. Load data:
    Layer \(\to\) Add Layer \(\to\) Add Vector Layer...

Click on Source \(\to\) [...]

Navigate to folder with file, select mi_2020.geojson, click Open

Click Add

You should see a map of Michigan’s electoral precincts

You can explore the vote tallies in each precinct by using the Identify Features tool (looks like an “i” in a circle). Using this tool, click on any polygon.

  1. Color precincts by vote share:
    right-click on layer’s name (mi_2020), select Properties...

Select Symbology type \(\to\) Graduated

Set graphical parameters (select variable):
Value \(\to\) VOTESHARE (percent Trump)

Set graphical parameters (select colors):
Color ramp \(\to\) All Color Ramps \(\to\) RdBu

Set graphical parameters (select colors):
Color ramp \(\to\) Invert Color Ramps (flip blue and red)

Set graphical parameters (make borders transparent):
Symbol \(\to\) Simple fill \(\to\) Stroke color \(\to\) Transparent stroke

Set graphical parameters (select interval type):
Mode \(\to\) Equal interval

Set graphical parameters (center intervals at 50%):
Symmetric classification \(\to\) Around \(=\) 50.0.

Click Classify, then OK.

The precincts should now be colored on a blue - red gradient.

  1. Save the map as an image
    Project \(\to\) New Print Layout...

Name the print layout “Map 1”

Add map:
Add item \(\to\) Add map

Add map:
use mouse (click and drag) to place map on layout

Add legend:
Add item \(\to\) Add legend

Add legend:
a legend should appear, with subtitle mi_2020

Add legend (change legend title):
Item Properties \(\to\) Title \(=\) “Trump’s 2020 vote share (%)”

Add scale bar:
Add item \(\to\) Add scale bar

Add scale bar: a scale bar should appear, units are metric by default

Add scale bar (change units to miles):
Item Properties \(\to\) Units \(\to\) Scalebar units \(=\) “Miles”

Save image to file:
Layout \(\to\) Export as image...

Select folder, name the file mi_2020_voteshare.png, click Save

Accept default settings, click Save

The image file should look something like this. Not pretty, but it’s the first pancake.

  1. Create new variable: vote margin (percent Trump \(-\) percent Biden)
    Right click mi_2020 layer \(\to\) Open attribute table

On top of spreadsheet, click Open field calculator button

  • Create New Field \(\to\) Output field name \(=\) VOTEMARGIN
  • Output field type \(=\) “Decimal number (real)”

Expression: (G20PRERTRU-G20PREDBID)/(G20PRERTRU+G20PREDBID)*100
Click OK

Repeat the same steps as before:
Color precincts by vote margin, symmetric classification Around \(=\) 0.00

Repeat the same steps as before:
Save map as image (mi_2020_votemargin.png)

Optional: to make middle color purple (not white), go to Color ramp \(\to\) Create New Color Ramp... in the Symbology tab

Select Gradient as color ramp type

This will open a new screen, where you can select the colors you want

Click on Color 1 and select a blue color

Click on Color 2 and select a red color

The middle color should now be purple. Click OK

The color classifications in the Symbology tab should automatically update. Click OK again

Closely-contested precincts should now appear as purple, not white

Problem Set

Can you make this map?

Your assignment (if using QGIS):
create the same kind of map for Massachusetts, using dataset ma_2020.geojson

  • precincts colored by VOTEMARGIN (new variable)
  • color ramp could be blue-white-red (shown here) or blue-purple-red
  • name the file ma_2020_votemargin.png
  • upload map to Canvas (by next Wednesday)

R

Introduction and Demo

Quick introduction to R

 

R is a programming language, which means you need to type commands, rather than pointing-and-clicking (unlike QGIS)

 

R scripts are plain text files with extension .R

  • to create a new script in RStudio, go to
    File menu \(\to\) New File \(\to\) R Script
  • …or open an existing script with File \(\to\) Open File...
  • save your script with File \(\to\) Save or File \(\to\) Save As...
    (remember to save it with extension .R)

 

The advantage of running scripts is that

  • it’s faster (once you get the hang of it),
  • more efficient, and
  • you get the same result every time

To run a single line of code, type a command, place your cursor at the line and hit Ctrl+Enter (Windows, Linux) or Command+Enter (MacOS)

print("Hello world!")
## [1] "Hello world!"

 

Let’s try some arithmetic

2+2
## [1] 4
2-2
## [1] 0
2*2
## [1] 4
2/2
## [1] 1

Basic syntax

 

To create a vector, we use

  • assignment operators = or <-
  • grouping operator c()
x = c(1,2,3,4,5)
x
## [1] 1 2 3 4 5
x <- c(1,2,3,4,5) # Equivalent to above
x
## [1] 1 2 3 4 5
y = x^2
y
## [1]  1  4  9 16 25

Basic syntax (cont’d)

 

Combine vectors into a dataset with the data.frame() command

my_dataset = data.frame(x=x,y=y)
summary(my_dataset)
##        x           y     
##  Min.   :1   Min.   : 1  
##  1st Qu.:2   1st Qu.: 4  
##  Median :3   Median : 9  
##  Mean   :3   Mean   :11  
##  3rd Qu.:4   3rd Qu.:16  
##  Max.   :5   Max.   :25

Basic plotting

 

Create a basic scatterplot with the plot() function

plot(x,y)

Getting help

 

To access the help file for any function, type ? before the function’s name

?c
?data.frame
?summary
?plot

There is also an abundance of online help forums for R users:

Loading external packages

 

For this exercise, we will need two external packages:

  • sf (“simple features”, to handle spatial vector data)
  • PlotTools (to customize plot legends)

 

We install packages with install.packages() and load them with library()

install.packages("sf") # (first time only)
install.packages("PlotTools") # (first time only)
library(sf)
library(PlotTools)

Setting working directory

 

We should always set a working directory, using the setwd() function.

The working directory is a folder on your hard drive, where you stored the data for this exercise, and where you’ll be saving maps.

 

You can find the path of a folder through Right-click \(\to\) Get Info (MacOS) or Right-click \(\to\) Properties (Windows).

 

On MacOS and Linux:

setwd("/home/username/Documents/SEST6577")

On Windows:

setwd("C:/Documents/SEST6577")

To find out your current working directory:

getwd()

Loading spatial vector data

 

Load the data through the read_sf() function from the sf package

mi_2020 = sf::read_sf(dsn = "Data/mi_2020.geojson")

Let’s take a look at the first 5 rows of mi_2020, using print(mi_2020,n=5)

print(mi_2020,n=5)
## Simple feature collection with 4751 features and 16 fields
## Geometry type: MULTIPOLYGON
## Dimension:     XY
## Bounding box:  xmin: -90.41829 ymin: 41.69613 xmax: -82.41348 ymax: 48.26269
## Geodetic CRS:  WGS 84
## # A tibble: 4,751 × 17
##   PRECINCTID      COUNTYFIPS cousubname elexpre G20PRERTRU G20PREDBID G20PRELJOR
##   <chr>           <chr>      <chr>      <chr>        <dbl>      <dbl>      <dbl>
## 1 WP-001-01040-0… 001        Alcona to… 001-AL…        564        248          3
## 2 WP-001-12460-0… 001        Caledonia… 001-CA…        508        245          4
## 3 WP-001-19320-0… 001        Curtis to… 001-CU…        486        238          2
## 4 WP-001-34820-0… 001        Greenbush… 001-GR…        560        302          9
## 5 WP-001-35740-0… 001        Gustin to… 001-GU…        317        112          9
## # ℹ 4,746 more rows
## # ℹ 10 more variables: G20PREGHAW <dbl>, G20PRENDEL <dbl>, G20PRETBLA <dbl>,
## #   G20USSRJAM <dbl>, G20USSDPET <dbl>, G20USSGSQU <dbl>, G20USSNDER <dbl>,
## #   G20USSTWIL <dbl>, VOTESHARE <dbl>, geometry <MULTIPOLYGON [°]>

This tells us:

  • the object class (Simple feature collection)
  • the object’s geometry type (MULTIPOLYGON) and spatial extent
  • its projection attributes (WGS 84)… more on this later
  • and a preview of the attribute table (A tibble: 4,751 x 17)

Visualizing spatial vector data

 

Let’s visualize the study region with plot(mi_2020["geometry"])

plot(mi_2020["geometry"])

Now let’s plot some attributes…

For a categorical variable (win/lose), visualization is simple…

 

  1. Create a vector of colors, where each precinct Trump won is coded red and every each precinct won by Biden is blue
cols = ifelse(mi_2020$G20PRERTRU>mi_2020$G20PREDBID,"red","blue")
plot(mi_2020["geometry"],col=cols)
  1. Use the resulting color vector with the plot(col=...) command.
plot(mi_2020["geometry"],col=cols)

 

 

Change the border’s thickness by changing lwd parameter

plot(mi_2020["geometry"],col=cols,lwd=.1)

Make border semi-transparent through alpha parameter in rgb

plot(mi_2020["geometry"],col=cols,lwd=.1,
    border=rgb(red = 0,green = 0,blue = 0,alpha = .5))

  1. Add a map legend with legend() command
legend(
  x = "bottomleft",
  title = "2020 Election Results",
  fill = c("red","blue"),
  legend =c("Trump win","Biden win")
)

Visualizing a continuous variable (e.g. vote share) is a little more tricky.

 

A relatively simple approach is to accept the default color palette like this:

plot(mi_2020["VOTESHARE"])

Or we can create a custom color palette (e.g. a red - blue gradient).

 

  1. Let’s set the number of breaks (e.g. 100)
breaks = seq(0,100,by=1)
  1. Create an RGB color ramp for the VOTESHARE variable
ramp = findInterval(mi_2020$VOTESHARE,breaks)
cols = rgb(ramp/length(breaks),0,(length(breaks)-ramp)/length(breaks))
  1. Use this vector of colors in the plot(col=...) command
plot(mi_2020["VOTESHARE"],col=cols)

Add a title to the plot, with main parameter

plot(mi_2020["VOTESHARE"],col=cols,
    main="Michigan 2020 Presidential Election Results")

Add a gradient legend, with the SpectrumLegend command from PlotTools

PlotTools::SpectrumLegend(
  x = "bottomleft", title = "Donald Trump's\n vote share (%)",
  palette = rgb(red=(100:0)/100,green=0,blue=(0:100)/100),
  legend = seq(0,100,by=20), lwd = 10
)

Creating a new variable in R is quite simple

mi_2020$VOTEMARGIN = 100*(mi_2020$G20PRERTRU-mi_2020$G20PREDBID)/
                        (mi_2020$G20PRERTRU+mi_2020$G20PREDBID)

where

  • mi_2020 is the dataset, and $ calls variables within it
  • to the left of the = sign is the new variable you want to create
  • to the right of = is an expression, specifying the calculation you want to make: \(\text{Vote Margin}= 100\cdot \frac{\text{Trump}-\text{Biden}}{\text{Trump}+\text{Biden}}\)

To plot VOTEMARGIN, we can follow the same procedure as before, just modifying the color breaks and legend to reflect the new range of this variable

breaks = seq(-100,100,by=1)
...
PlotTools::SpectrumLegend(..., legend = seq(-100,100,by=25), ...)

To export your map as an image file, use the png() command
(or pdf(), jpeg(), tiff() for other file formats)

png(filename="Output/mi_2020_votemargin_r.png",width=6,height=6,units="in",res=150)
par(mar=c(0,0,0,0))
plot(mi_2020["VOTEMARGIN"],col=cols,lwd=.1,border=rgb(0,0,0,.1),
    main="Michigan 2020 Presidential Election Results")
PlotTools::SpectrumLegend(x = "bottomleft",title = "Donald Trump's\n vote margin (%)",
    palette = rgb(red=(100:0)/100,green=0,blue=(0:100)/100), 
    legend = seq(-100,100,by=20), lwd = 20, bty = "n")
dev.off()

where

  • the filename argument specifies the location (e.g. Output/) and name (e.g. mi_2020_votemargin_r.png), in one string
  • the width and height arguments specify the dimensions of the file
  • units specifies the units of measurement (in \(=\) inches) and res specifies the resolution (150 pixels per inch)
  • dev.off() closes the graphics device
    (everything between png() and dev.off() is printed to file)

 

 

 

 

The exported image should look like this: \(\to\)


mi_2020_votemargin_r.png

Problem Set

 

 

Your assignment (if using R):
create the same kind of map for Massachusetts, using dataset ma_2020.geojson

  • precincts colored by VOTEMARGIN (new variable)
  • name the file ma_2020_votemargin_r.png
  • upload map to Canvas
    (by next Wednesday)


Can you make this map?