Goal of lab exercise: Make multi-layer maps

 

  1. Lab exercise
    1. Map of violence and terrain in Afghanistan
    2. Map of violence and ethnicity in Afghanistan
  2. Problem set
    1. Map of violence and ethnicity in Yemen
      • same approach as the second map above \(\uparrow\)
      • name the file yem_map_2.png
      • upload map to Canvas (by next Wednesday)

Violence and terrain in Afghanistan (we’ll make this)

Violence and ethnicity in Afghanistan (this too)

Violence and ethnicity in Yemen (you’ll make this)

You can make these plots in QGIS or in R. Instructions for both are below.

Map 1 in R


Map 2 in R


PS02 in R

We will put multiple types of data on same map:

Category Type Format Data source
Elevation (meters) Raster .tif ETOPO
Province borders Vector (polygon) .shp GADM
Roads Vector (polyline) .shp DCW
Violence (1989-2016) Vector (point) .csv xSub/UCDP
Ethnic settlement patterns Vector (polygon) .shp GeoEPR

 

These are all in the PS02.zip file posted on Canvas.

 

Let’s open QGIS…

QGIS

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

Map 1

Load elevation data: Layer \(\to\) Add Layer \(\to\) Add Raster Layer...

Press the ... button to select the file

Navigate to ETOPO1_downsampled.tif file in Data/Elevation/ folder, click Open

Click Add on next screen and go back to main window.

You should now see a global elevation raster

Load Afghanistan borders: Layer \(\to\) Add Layer \(\to\) Add Vector Layer...
Click ... in dialog box to select the file.

Navigate to gadm36_AFG_1.geojson in Data/Borders/ folder.
Click Open, then Add

Afghanistan’s borders should appear on the map. But they are hard to see

Right click on gadm36_AFG_1 in the layer menu. Click Zoom to Layer

Now we can see Afghanistan, but the elevation layer is hidden behind it

Make polygons transparent: Right-click on gadm36_AFG_1 in layer menu, click on Properties.  Highlight Simple fill in the symbology menu.

Select Fill style: No Brush

Click OK

The polygons should now be transparent, with provincial borders still visible.

Let’s drop the rest of the raster, keeping only the overlap with Afghanistan.
Clip raster: Raster \(\to\) Extraction \(\to\) Clip Raster by Extent...

Set the Input layer as ETOPO1_downsampled.
Click Clipping extent \(\to\) Calculate from Layer \(\to\) gadm36_AFG_1

Click Run

Check log to make sure there were no errors.

The output is a temporary raster dataset, clipped to extent of Afghanistan.

Rename the new layer to “Elevation”. Right click Clipped \(\to\) Rename Layer

Change the name to Elevation

Let’s do the same for gadm36_AFG_1

Change its name to Borders.
Also, un-check the box next to ETOPO1_downsampled to hide the big raster

Change plot order. Drag Borders above Elevation in Layer menu.
Borders should now be (faintly) visible above elevation layer.

Change raster color. Right click Elevation \(\to\) Properties.
Click on the menu next to Render type

Select Render type: Singleband pseudocolor

Click menu button next to Color ramp \(\to\) Create New Color Ramp...

Select Catalog: cpt-city as color ramp type

This will open a window with a lot of color themes.
Select Topography \(\to\) elevation. Click OK

Click Classify button, then OK

The elevation layer should now appear in blue-green-brown

Add roads layer. Same Add Layer procedure as for borders. Open AFG_roads.geojson from Data/Roads/ folder

You should see the AFG_roads layer on the map

Change layer’s name from AFG_roads to Roads

Change color to blue. Right click on Roads \(\to\) Properties.
Highlight Simple line in Symbology menu

Click menu button next to Color. Select blue from Standard colors list

Click Apply or OK

The roads should appear in blue

Add violence data. Layer \(\to\) Add Layer \(\to\) Add Delimited Text Layer...

Click ... button next to File name dialog

Navigate to xSub_GED_AFG_event.csv in Data/Violence/ menu.
Click Open

Under Geometry Definition, set X field \(=\) LONG and Y field \(=\) LAT.
Click Add

You should now see about 30,000 violent events on the map

Change layer’s name to Violence




Change layer’s color. Right click \(\to\) Properties. Click on dot white (\(\circ\))

Make points smaller and translucent. Set Opacity \(=\) 50% and Size \(=\) 2.0

The rendered points should look something like this.

Export the map to image. Project \(\to\) New Print Layout.
Place map on layout, with Add Item \(\to\) Add Map

Add legend with Add Item \(\to\) Add Legend.
The legend includes an item we need to remove (ETOPO1_downsampled)

To remove this item, highlight the legend, go to Item Properties. Un-check the Auto update box under Legend Items. Highlight ETOPO1_downsampled and click on the red “\(-\)” button to remove it

The ETOPO1_downsampled item should now disappear from the legend.

Edit layers’ names in legend. Highlight Elevation and click on the Edit button (pencil and paper icon). Change text to “Elevation (m)”




Let’s also remove the band information from the legend. Highlight Band 1: ETOPO1_Bed..., click the “–” button to remove it.

Change legend Title to “Violence in Afghanistan” in Item Properties

The legend should now look something like this

Add scale bar. Change Scalebar units to Miles

Export to file. Layout menu \(\to\) Export as Image.... Name it afg_map_1.png

The output file should look like this.

Map 2

Load ethnicity data. Add (as vector layer) GeoEPR_AFG.geojson from Data/Ethnicity/ folder

This will add a new polygon layer to the map.
Change its name from GeoEPR_AFG to Ethnicity

Open the layer’s Properties. Click on Single symbol in Symbology menu

Change symbol type from Single symbol to Categorized

Make polygon boundaries transparent. Click on menu button next to Stroke color \(\to\) check Transparent Stroke box

Select Value \(=\) group

Click on Classify button, then Apply or OK

Each ethnic group’s polygon should appear as a different color

Create a new map layout and export it as afg_map_2.png. You can do it

Problem Set 2

 

Your assignment (if using QGIS):

  • create the same kind of “violence and ethnicity” map for Yemen
  • use these files:
    • GeoEPR_YEM.geojson
    • xSub_GED_YEM_event.csv
    • gadm36_YEM_1.geojson
  • follow the same steps as for “Map 2” above
  • name it yem_map_2.png
  • upload map to Canvas
    (by next Wednesday)


Can you make this map?

R

Loading R packages

 

To implement these steps in R, we will be using two packages:

  • sf (“simple features”, to handle spatial vector data)
  • terra (to handle raster data, like Elevation)
library(sf)
library(terra) # NEW!

NOTE: The code to produce Map 1 and Map 2 in R is in ps02_demo.R on RStudio Cloud, and in PS02.zip (posted on Canvas).

Map 1

Loading spatial data

 

Let’s load the elevation raster into R, using the rast() command from terra:

elevation_global = terra::rast("Data/Elevation/ETOPO1_downsampled.tif")
plot(elevation_global)

Let’s load the provincial boundaries and roads into R, using sf::read_sf():

borders = sf::read_sf("Data/Borders/gadm36_AFG_1.geojson")
roads = sf::read_sf("Data/Roads/AFG_roads.geojson")
plot(borders["geometry"])
plot(roads["geometry"])

Let’s load the violence data into R, using read.csv():

violence = read.csv("Data/Violence/xSub_GED_AFG_event.csv")

To convert this table into a spatial object, we will use the sf::st_as_sf() command, specifying columns corresponding to geographic coordinates (LONG, LAT), and copying the CRS from the borders object:

violence = sf::st_as_sf(violence, coords=c("LONG","LAT"))
sf::st_crs(violence) = sf::st_crs(borders)
plot(violence["geometry"])

Clipping the raster to Afghanistan’s extent

 

R’s equivalent of QGIS’s Clip Raster by Extent... command is the following:

  1. Create a rectangular polygon corresponding to the extent of the borders
    • sf::st_bbox() finds the spatial extent of a layer
    • sf::st_as_sfc() converts the extent object into a polygon
extent_afg = sf::st_as_sfc(sf::st_bbox(borders))
  1. Crop the raster by this new polygon, using terra::crop()
elevation = terra::crop(elevation_global, extent_afg)
plot(elevation, col=terra::map.pal("elevation")); 
plot(borders["geometry"], add=TRUE)

Plotting Map 1

 

To create the map, we will execute the following commands in sequence:

  • set margins, par()
  • draw elevation layer, plot() (note different syntax for rasters vs. vectors)
  • add violence (transluscent dark red points)
  • add borders, roads (adjusting thickness, color)
  • add legend() (specify x and y coordinates for top-left corner of legend)
par(mar=c(2,2,.5,.5))
plot(elevation, col=terra::map.pal("elevation"))
plot(violence["geometry"],
plot(roads["geometry"],add=T,col="black",lwd=1.618)
plot(roads["geometry"],add=T,col="pink",lwd=.62)
plot(borders["geometry"],add=T,lwd=2)
legend(x=70.5,y=31.5,legend=c("Violence","Borders","Roads"),
  pch=c(1,NA,NA),lwd=c(NA,2,1),title="Violence in Afghanistan",
  bg="white",col=c("black","black","pink4"))

 

Here’s what the plot should look like:

Exporting Map 1 to image file

 

To save the map, we will use the png() and dev.off() commands, as in PS01:

png("Output/afg_map_1_r.png",width = 8,height = 6,
  units = "in",res = 300)
par(mar=c(2,2,.5,.5))
plot(elevation, col=terra::map.pal("elevation"))
plot(violence["geometry"],
plot(roads["geometry"],add=T,col="black",lwd=1.618)
plot(roads["geometry"],add=T,col="pink",lwd=.62)
plot(borders["geometry"],add=T,lwd=2)
legend(x=70.5,y=31.5,legend=c("Violence","Borders","Roads"),
  pch=c(1,NA,NA),lwd=c(NA,2,1),title="Violence in Afghanistan",
  bg="white",col=c("black","black","pink4"))
dev.off()

The output file should look like this.

Map 2

Loading ethnicity data

 

Let’s load the ethnic settlements data into R, using sf::read_sf():

ethnicity = sf::read_sf("Data/Ethnicity/GeoEPR_AFG.geojson")
plot(ethnicity["geometry"])

Plotting Map 2

 

Let’s begin by creating a vector of colors, corresponding to the group column:

groups = as.factor(ethnicity$group)
n = length(unique(groups))
cols = hcl.colors(n,palette="Pastel 1")[groups]
plot(ethnicity["group"],col=cols,border="white")

To see what other color palettes are available in hcl.colors(), see hcl.pals():

hcl.pals()
##   [1] "Pastel 1"      "Dark 2"        "Dark 3"        "Set 2"        
##   [5] "Set 3"         "Warm"          "Cold"          "Harmonic"     
##   [9] "Dynamic"       "Grays"         "Light Grays"   "Blues 2"      
##  [13] "Blues 3"       "Purples 2"     "Purples 3"     "Reds 2"       
##  [17] "Reds 3"        "Greens 2"      "Greens 3"      "Oslo"         
##  [21] "Purple-Blue"   "Red-Purple"    "Red-Blue"      "Purple-Orange"
##  [25] "Purple-Yellow" "Blue-Yellow"   "Green-Yellow"  "Red-Yellow"   
##  [29] "Heat"          "Heat 2"        "Terrain"       "Terrain 2"    
##  [33] "Viridis"       "Plasma"        "Inferno"       "Rocket"       
##  [37] "Mako"          "Dark Mint"     "Mint"          "BluGrn"       
##  [41] "Teal"          "TealGrn"       "Emrld"         "BluYl"        
##  [45] "ag_GrnYl"      "Peach"         "PinkYl"        "Burg"         
##  [49] "BurgYl"        "RedOr"         "OrYel"         "Purp"         
##  [53] "PurpOr"        "Sunset"        "Magenta"       "SunsetDark"   
##  [57] "ag_Sunset"     "BrwnYl"        "YlOrRd"        "YlOrBr"       
##  [61] "OrRd"          "Oranges"       "YlGn"          "YlGnBu"       
##  [65] "Reds"          "RdPu"          "PuRd"          "Purples"      
##  [69] "PuBuGn"        "PuBu"          "Greens"        "BuGn"         
##  [73] "GnBu"          "BuPu"          "Blues"         "Lajolla"      
##  [77] "Turku"         "Hawaii"        "Batlow"        "Blue-Red"     
##  [81] "Blue-Red 2"    "Blue-Red 3"    "Red-Green"     "Purple-Green" 
##  [85] "Purple-Brown"  "Green-Brown"   "Blue-Yellow 2" "Blue-Yellow 3"
##  [89] "Green-Orange"  "Cyan-Magenta"  "Tropic"        "Broc"         
##  [93] "Cork"          "Vik"           "Berlin"        "Lisbon"       
##  [97] "Tofino"        "ArmyRose"      "Earth"         "Fall"         
## [101] "Geyser"        "TealRose"      "Temps"         "PuOr"         
## [105] "RdBu"          "RdGy"          "PiYG"          "PRGn"         
## [109] "BrBG"          "RdYlBu"        "RdYlGn"        "Spectral"     
## [113] "Zissou 1"      "Cividis"       "Roma"

To create the map, we will use the same commands as above, with some differences:

  • set margins to zero
  • include option reset=FALSE when plotting the ethnicity layer
  • add violence data twice (white points + black outlines)
  • add legend() (including the vector of group colors)
par(mar=c(0,0,0,0))
plot(ethnicity["group"],col=cols,border=NA,reset=FALSE,main="")
plot(borders["geometry"],add=T,lwd=2)
# Add violence (white points + black outlines)
plot(violence["geometry"],col=rgb(1,1,1,alpha = .25),pch=16,add=T)
plot(violence["geometry"],col=rgb(0,0,0,alpha = .25),pch=1,add=T)
# Add legend
legend("bottomright",bty="n",ncol=2,
  legend=c("Violence","Borders",as.character(groups)),
  col=c("black","black",cols),pch=c(1,0,rep(15,length(cols))),
  cex=.8,title="Ethnicity and violence\n in Afghanistan"
  )

 

Here’s what the plot should look like:

Exporting Map 2 to image file

 

Using the png() and dev.off() commands as bookends:

png("Output/afg_map_2_r.png",width = 8,height = 6,
  units = "in",res = 300)
par(mar=c(0,0,0,0))
plot(ethnicity["group"],col=cols,border=NA,reset=FALSE,main="")
plot(borders["geometry"],add=T,lwd=2)
# Add violence (white points + black outlines)
plot(violence["geometry"],col=rgb(1,1,1,alpha = .25),pch=16,add=T)
plot(violence["geometry"],col=rgb(0,0,0,alpha = .25),pch=1,add=T)
# Add legend
legend("bottomright",bty="n",ncol=2,
  legend=c("Violence","Borders",as.character(groups)),
  col=c("black","black",cols),pch=c(1,0,rep(15,length(cols))),
  cex=.8,title="Ethnicity and violence\n in Afghanistan")
dev.off()

The output file should look like this.

Problem Set 2

 

Your assignment (if using R):

  • create the same kind of “violence and ethnicity” map for Yemen
  • use these files:
    • GeoEPR_YEM.geojson
    • xSub_GED_YEM_event.csv
    • gadm36_YEM_1.geojson
  • follow the same steps as for “Map 2” here
  • name it yem_map_2_r.png
  • upload map to Canvas
    (by next Wednesday)


Can you make this map?