plotting points with geom_sf (simple features format)

[GreenStat]

copied from r-spatial/sf#88

HI, I tried point and line geometry with (the experimental) 'geom_sf' for plotting maps (in simple features format ):

library(sf)
library(maps)
library(ggplot2)

#points
data(meuse,package="sp")
meuse_sf <- st_as_sf(meuse, coords = c("x", "y"), crs = 28992, agr = "constant")

summary(meuse_sf)

meuse_sf %>% ggplot() + geom_sf(aes(size=cadmium))

this gives:

I am a bit confused by the behaviour of aes in geom_sf, with aes(size=cadmium) I expected varying size of dots, instead it gives me varying size of line thickness of the circles.
How do I change type and size of dots?

meuse_sf %>% ggplot() + geom_sf(aes(color=cadmium)) gives:

line color of the circles changes, I want a fill....

meuse_sf %>% ggplot() + geom_sf(aes(fill=cadmium))
this gives:

more confusing, I still want a fill.... :-), how can I do that ?

[hadley]

Could you help me about by making a very simple point dataset that I could include in a unit test?

[GreenStat]

sure,
maybe you can give me a pointer to the type of dataset you need. I could tweek the meuse data by adding factor, posix(?), char column, is that the kind of file you need ?

[hadley]

It's more the opposite - I need a few lines of code to that create a dataset with ~4 points without any external dependencies

[GreenStat]

does this works ? (I am not very familiar with dput)

dput(simple.sf)
 
structure(list(names = c("utrecht", "groningen", "amsterdam", 
"munster"), geometry = structure(list(`1` = structure(c(5.1214201, 
52.0907374), class = c("XY", "POINT", "sfg")), `2` = structure(c(6.5665018, 
53.2193835), class = c("XY", "POINT", "sfg")), `3` = structure(c(4.8951679, 
52.3702157), class = c("XY", "POINT", "sfg")), `4` = structure(c(7.6261347, 
51.9606649), class = c("XY", "POINT", "sfg"))), .Names = c("1", 
"2", "3", "4"), n_empty = 0L, precision = 0, crs = structure(list(
    epsg = 4326L, proj4string = "+proj=longlat +datum=WGS84 +no_defs"), .Names = c("epsg", 
"proj4string"), class = "crs"), bbox = structure(c(4.8951679, 
51.9606649, 7.6261347, 53.2193835), .Names = c("xmin", "ymin", 
"xmax", "ymax")), class = c("sfc_POINT", "sfc"))), .Names = c("names", 
"geometry"), row.names = c(NA, -4L), sf_column = "geometry", agr = structure(NA_integer_, class = "factor", .Label = c("constant", 
"aggregate", "identity"), .Names = "names"), class = c("sf", 
"data.frame"))

[GreenStat]

A simple script version:

lon<-c(5.121420, 6.566502, 4.895168, 7.626135)
lat<-c(52.09074, 53.21938, 52.37022, 51.96066)
cities<-c('utrecht','groningen','amsterdam','munster')
size<-c(300,500,1000,50)

xy.cities<-data.frame(lon,lat,cities,size)
simple.sf <- st_as_sf(xy.cities, coords=c('lon','lat'))
st_crs(simple.sf) <- 4326
simple.sf %>% ggplot() + geom_sf(aes(size=size))

[hadley]

Yeah, the script version is much better, thanks.

[hadley]

Here's a minimal reprex:

cities <- tibble::tribble(
  ~ lon,    ~ lat,    ~ name,      ~ pop,
  5.121420, 52.09074, "Utrecht",   311367,
  6.566502, 53.21938, "Groningen", 189991,
  4.895168, 52.37022, "Amsterdam", 779808 
) %>% sf::st_as_sf(coords = c("lon", "lat"), crs = 4326)

lines_sfc <- sf::st_sfc(list(
  sf::st_linestring(rbind(cities$geometry[[1]], cities$geometry[[2]])),
  sf::st_linestring(rbind(cities$geometry[[2]], cities$geometry[[3]]))
))

lines <- sf::st_sf(
  id = 1:2,
  size = c(10,50),
  geometry = lines_sfc,
  crs = 4326
)

ggplot() + 
  geom_sf(aes(size = pop), data = cities) + 
  geom_sf(aes(colour = size), data = lines)

The problem is that the geom needs to learn what legend to use from the data, but there's currently no way to pass this information between the two function calls. The documentation for Geom notes:

Compared to \code{Stat} and \code{Position}, \code{Geom} is a little
different because the execution of the setup and compute functions is
split up. \code{setup_data} runs before position adjustments, and
\code{draw_layer} is not run until render time, much later. This
means there is no \code{setup_params} because it's hard to communicate
the changes.

[boshek]

Some additional context as well. It looks like shape is not currently supported by geom_sf:

> ggplot(cities) + 
+   geom_sf(shape = 5)
Warning: Ignoring unknown parameters: shape