X-Git-Url: http://git.vanrenterghem.biz/R/project-wa-schools.git/blobdiff_plain/f27fc3a28ef7ba08696e0075191311dfccfca458..4a262294342cd273e52b82bb33d371904f840449:/WA_active_schools.R

diff --git a/WA_active_schools.R b/WA_active_schools.R
index 7792496..5076c60 100644
--- a/WA_active_schools.R
+++ b/WA_active_schools.R
@@ -1,17 +1,20 @@
 # install libgdal-dev on debian first
-#install.packages("rgdal")
-library(rgdal)
+# install libudunits2-dev on debian first
+#install.packages("sf")
+library(sf)
 library(tidyverse)
+library(broom)
 #install.packages("OpenStreetMap")
 library(OpenStreetMap)
 library(viridis)
 
 # Read the shapefile obtained from WA opendata (SLIP)
-shape <- readOGR(dsn = "data/", layer = "CurrentActiveSchoolsSemester12017_PublicDET_014")
-shape.df <- tidy(shape)
-shape.df$latitude <- as.numeric(as.character(shape.df$latitude))
-shape.df$longitude <- as.numeric(as.character(shape.df$longitude))
-shape.df$totalschoo <- as.numeric(as.character(shape.df$totalschoo))
+#shape <- readOGR(dsn = "data/", layer = "CurrentActiveSchoolsSemester12017_PublicDET_014")
+shape.df <- st_read(dsn = "data/", layer = "CurrentActiveSchoolsSemester12017_PublicDET_014", stringsAsFactors = FALSE)
+shape.df$latitude <- as.numeric(shape.df$latitude)
+shape.df$longitude <- as.numeric(shape.df$longitude)
+shape.df$totalschoo <- as.numeric(shape.df$totalschoo)
+shape.df$totalsecon <- as.numeric(shape.df$totalsecon)
 
 # filter out remote islands
 shape.df <- filter(shape.df,longitude > 100)
@@ -62,19 +65,29 @@ shape.metro.df <- filter(shape.df, abs(latitude) <= max(abs(lat.metro)),
                          abs(longitude) <= max(abs(lon.metro)),
                          abs(longitude) >= min(abs(lon.metro)))
 mapLatLon.metro <- openproj(map.metro)
+# Let's look at secondary school only
+# 95th percentile of school size
+secon95th <- quantile(shape.metro.df[,"totalsecon">0]$totalsecon,probs = .95)
 autoplot(mapLatLon.metro) +
-  geom_point(data = shape.metro.df, 
+  geom_point(data = filter(shape.metro.df, totalsecon > 0), # geom_sf is not yet in ggplot2 on CRAN, so sticking to this for now
              aes(x = longitude, 
                  y = latitude,
-                 color = totalschoo),
-             show.legend = TRUE) +
+                 color = totalsecon,
+                 size = totalsecon)
+             ) +
+  geom_text(data = filter(shape.metro.df, totalsecon > secon95th), # name top 5% schools
+             aes(x = longitude, 
+                 y = latitude,
+                 label = schoolname),
+            size = 2) +
   scale_color_viridis(option="viridis",
                       name = "Number\nof pupils") +
+  guides(size = FALSE) + # Do not show legend for size, as color is the main key
   theme(legend.position = "right",
         axis.text = element_blank(),
         axis.ticks = element_blank()) +
-  labs(title = "Western Australian schools",
-       subtitle = "Perth metro area only",
+  labs(title = "Western Australian secondary schools",
+       subtitle = "Perth metro area",
        caption = "Map tiles by Stamen Design, under CC BY 3.0.\nMapdata by OpenStreetMap, under ODbL.
        School data © Government of WA 2017.\n Dep. of Education & Training.",
        x="",