Article 7: Visualizing with maps

2022-06-09

Source: vignettes/a07_visual_maps.Rmd

While ALUES can generate tables for suitability scores and classes, it can also take advantage of the rich mapping library of R. The main requirement of course is the availability of the longitude and latitude for each of the land units. This is possible for Marinduque as it has spatial variables.

Suitability scores and classes

Suppose we want to evaluate the land units for banana, then:

## Loading required package: Rcpp
y <- MarinduqueLT
banana_suit <- suit("banana", terrain=y)
## Warning in suitability(terrain, crop_soil, mf = mf, sow_month = NULL, minimum
## = minimum, : maximum is set to 16 for factor CECc since all parameter intervals
## are equal.
banana_ovsuit <- overall_suit(banana_suit[["soil"]], method="average")

Generate maps

There are several ways to generate maps in R, but the following uses ggmap library:

library(ggmap)
library(raster)
library(reshape2)
map_lvl0 <- getData("GADM", country = "PHL", level = 0)
map_lvl2 <- getData("GADM", country = "PHL", level = 2)

prov <- map_lvl2[map_lvl2$NAME_1 == as.character("Marinduque"),]
munic_coord <- coordinates(prov)
munic_coord <- data.frame(munic_coord)
munic_coord$label <- prov@data$NAME_2

val <- banana_suit[["soil"]][[2]]
val["Overall Suitability"] <- banana_ovsuit[,1]
d_map <- melt(as.matrix(val))
d_map$Lon <- rep(y$Lon, ncol(val)); d_map$Lat <- rep(y$Lat, ncol(val))

fill <- "#FFF7BC"; shadow <- "#9ECAE1"; ncol <- 3; size <- 3; alpha <- 1
text_opts <- list(alpha = 1, angle = 0, colour = "black", family = "sans", fontface = 1, lineheight = 1, size = 3)
labels <- list(title = "", xlab = "", ylab = "")

p1 <- ggplot() + geom_polygon(data = prov, aes(long + 0.008, lat - 0.005, group = group), fill = shadow) + 
  geom_polygon(data = prov, aes(long, lat, group = group), colour = "grey50", fill = fill) +
  geom_tile(aes(x = Lat, y = Lon, fill = value), data = d_map, size = size, alpha = alpha) +
  facet_wrap(~ Var2, ncol = ncol) +
  geom_polygon(data = prov, aes(long, lat, group = group), colour = "#4E4E4C", alpha = 0) +
  geom_label(data = munic_coord, aes(x = X1, y = X2, label = label), alpha = 0.5,
            angle = text_opts$angle, colour = "white", fill = "black", family = text_opts$family,
            fontface = text_opts$fontface,
            lineheight = text_opts$lineheight, size = text_opts$size) +
  coord_equal() + ggtitle(as.character(labels$title)) + xlab(as.character(labels$xlab)) + ylab(as.character(labels$ylab)) +
  scale_fill_gradientn(name = "Score\n", colors = c("red", "#FFDF00")) +
  scale_x_continuous(breaks = round(seq(min(d_map$Lat) + 0.05, max(d_map$Lat), len = 3), 2)) +
  theme(panel.background = element_rect(fill = '#F7E7CE'), 
        strip.background = element_rect(fill = "#D4BF96"),
        strip.text.x = element_text(size = 12),
        axis.text.x = element_text(size=12), 
        legend.text=element_text(size=12),
        legend.title=element_text(size=12),
        axis.text.y = element_text(size=12), legend.position = c(0.85, 0.25))
p1

And for suitability classes:

val <- banana_suit[["soil"]][[3]]
val["Overall Suitability"] <- banana_ovsuit[,2]
d_map <- melt(as.matrix(val))
d_map$Lon <- rep(y$Lon, ncol(val)); d_map$Lat <- rep(y$Lat, ncol(val))

d_map$Class <- factor(d_map$value, levels=c("N", "S3", "S2", "S1"))

p1 <- ggplot() + geom_polygon(data = prov, aes(long + 0.008, lat - 0.005, group = group), fill = shadow) + 
  geom_polygon(data = prov, aes(long, lat, group = group), colour = "grey50", fill = fill) +
  geom_tile(aes(x = Lat, y = Lon, fill = Class), data = d_map, size = size, alpha = alpha) +
  facet_wrap(~ Var2, ncol = ncol) +
  geom_polygon(data = prov, aes(long, lat, group = group), colour = "#4E4E4C", alpha = 0) +
  geom_label(data = munic_coord, aes(x = X1, y = X2, label = label), alpha = 0.5,
             angle = text_opts$angle, colour = "white", fill = "black", family = text_opts$family,
             fontface = text_opts$fontface,
             lineheight = text_opts$lineheight, size = text_opts$size) +
  coord_equal() + ggtitle(as.character(labels$title)) + xlab(as.character(labels$xlab)) + ylab(as.character(labels$ylab)) +
  scale_colour_discrete(name = "Class\n", breaks=c("N", "S3", "S2", "S1"), labels=c("N", "S3", "S2", "S1")) +
  scale_x_continuous(breaks = round(seq(min(d_map$Lat) + 0.05, max(d_map$Lat), len = 3), 2)) +
  theme(panel.background = element_rect(fill = '#F7E7CE'), 
        strip.background = element_rect(fill = "#D4BF96"),
        strip.text.x = element_text(size = 12),
        axis.text.x = element_text(size=12), 
        legend.text=element_text(size=12),
        legend.title=element_text(size=12),
        axis.text.y = element_text(size=12), legend.position = c(0.85, 0.25))
p1