Difficulty combining geom_path and geom_text on ggplot map - r

Question

I'm trying to use geom_path and geom_text to create a labeled map, but it is not going well. I can use each geom_path and geom_text separately, but I can't seem to get them to work together. I think it has something to do with geom_polygon, but I'm not sure what.
Here is how I prepared my shapefile for mapping:
meck.hiv <- as(meck.hiv, "Spatial")
meck.hiv#data$seq_id <- seq(1:nrow(meck.hiv#data)) #create unique id for each polygon
meck.hiv#data$id <- rownames(meck.hiv#data)
meck.hivdf <- fortify(meck.hiv) #fortify the data
zipcodedf <- merge(meck.hivdf, meck.hiv#data,
by = "id")# merge the "fortified" data with the data from our spatial object
cnames <- aggregate(cbind(long, lat) ~ zip, data=zipcodedf, FUN=function(x)mean(range(x))) #get the names of our zipcode (and center the coordinates for labels)
With this, I was able to get the following maps:
With labels, no paths:
p2 <- ggplot(data = zipcodedf, aes(x = long, y = lat)) +
geom_polygon(aes(group=group, fill=hispanic.dis)) +
geom_text(data=cnames, aes(long, lat, label = zip), size=3, fontface='bold', color="black")+ #This put in zip code names
scale_fill_brewer(breaks=c(1, 2, 3, 4, 5, 6, 7), labels=c("<5", "5-10", "11-15", "16-20", "21-25", "26-30", "31+"), palette="Reds",
na.value="darkgrey") +
coord_equal() +
theme(panel.background= element_rect(color="black")) +
theme(axis.title = element_blank(), axis.text = element_blank()) +
labs(title = "Hispanic Population by Zip Code", fill="Hispanic Population (% of Total)")
Resulting map with labels but no path
Paths, but no labels:
p3 <- ggplot(data = zipcodedf, aes(x = long, y = lat, group = group, fill = hispanic.dis)) +
geom_polygon() +
geom_path(color = "black", size = 0.2)+ #Oddly, use of geom_path with the above leads to weird stuff, but we can customize map lines without labels here
scale_fill_brewer(breaks=c(1, 2, 3, 4, 5, 6, 7), labels=c("<5", "5-10", "11-15", "16-20", "21-25", "26-30", "31+"), palette="Reds",
na.value="darkgrey") +
coord_equal() +
theme(panel.background=element_blank())+
theme(panel.background= element_rect(color="black")) +
theme(axis.title = element_blank(), axis.text = element_blank()) +
labs(title = "Hispanic Population by Zip Code", fill="Hispanic Population(% of Total)")
Resulting path with paths but no labels
Trying to combine
p4 <- ggplot(data = zipcodedf, aes(x = long, y = lat)) +
geom_polygon(aes(group=group, fill=hispanic.dis)) +
geom_text(data=cnames, aes(long, lat, label = zip), size=3, fontface='bold', color="black")+
geom_path(color = "black", size = 0.2)+ #Oddly, use of geom_path with the above leads to weird stuff, but we can customize map lines without labels here
scale_fill_brewer(breaks=c(1, 2, 3, 4, 5, 6, 7), labels=c("<5", "5-10", "11-15", "16-20", "21-25", "26-30", "31+"), palette="Reds",
na.value="darkgrey") +
coord_equal() +
theme(panel.background=element_blank())+
theme(panel.background= element_rect(color="black")) +
theme(axis.title = element_blank(), axis.text = element_blank()) +
labs(title = "Hispanic Population by Zip Code", fill="Hispanic Population(% of Total)")
Resulting screwed up plot with paths and labels
My trouble-shooting suggests that there's something about how geom_path is interacting with geom_polygon combined with ggplot(), but if I move all the instructions into geom_polygon, path doesn't show up at all (although a plot will generate), and if I move all the instructions into ggplot(), I get an error: "Error in FUN(X[[i]], ...): object 'group' not found" (below).
p5 <- ggplot(data = zipcodedf, aes(x = long, y = lat, group = group, fill = hispanic.dis)) +
geom_polygon() +
geom_path(color = "black", size = 0.2)+
geom_text(data=cnames, mapping=aes(x=long, y=lat))+#Oddly, use of geom_path with the above leads to weird stuff, but we can customize map lines without labels here
scale_fill_brewer(breaks=c(1, 2, 3, 4, 5, 6, 7), labels=c("<5", "5-10", "11-15", "16-20", "21-25", "26-30", "31+"), palette="Reds",
na.value="darkgrey") +
coord_equal() +
theme(panel.background=element_blank())+
theme(panel.background= element_rect(color="black")) +
theme(axis.title = element_blank(), axis.text = element_blank()) +
labs(title = "Hispanic Population by Zip Code", fill="Hispanic Population(% of Total)")
Any help is appreciated. Please let me know if I need to provide more information on anything.

Answer 1

{ggplot2}'s layering and inheritance can be confusing at times. I hope the following helps. Also check out whether you really want an "additional" geom_path() layer. It feels like you want to depict/highlight the county boundaries. Please also note the use of color for setting the boundary(line) color for a layer in the example.
I have to emulate your problem, as you did not provide a reproducible example. The initial code is to get me the map data for North Carolina. I construct a cnames dataframe for the county names. You should be able to apply this to your zip-code problem at the county level ... :)
library(dplyr)
library(ggplot)
library(maps) # helper package to get some map data
library(mapdata) # helper for some map data
counties <- map_data("county") %>% filter(region == "north carolina")
head(counties)
yields a data frame
long lat group order region subregion
1 -79.53800 35.84424 1857 54915 north carolina alamance
2 -79.54372 35.89008 1857 54916 north carolina alamance
3 -79.53800 35.98175 1857 54917 north carolina alamance
4 -79.52081 36.23385 1857 54918 north carolina alamance
5 -79.26298 36.23385 1857 54919 north carolina alamance
6 -79.27444 35.90726 1857 54920 north carolina alamance
Plotting the counties only (note: this would be your zip-code level).
To explain the principle of layer inheritance, I "set" all parameters in the layer call (i.e. geom_polygon()):
ggplot() +
geom_polygon(data = counties
, aes(x = long, y = lat, group = group) # aesthetic mapping
, color = "white" # fixed value for "line"
, fill = "lightblue") # fixed value for "fill"
If you now would add the geom_path() layer without giving it aesthetics, the plot will not change. Check that I set the color to blue and the linesize to 2 for demo purposes.
ggplot() +
geom_polygon(data = counties, aes(x = long, y = lat, group = group), colour = "white", fill = "lightblue") +
# ---------------- adding a layer with no aesthetics -----------
geom_path(color = "blue", size = 2) # ... no joy!
If you now move the data and aesthetics to ggplot() "base"-layer, also path will inherit the aesthetics. In this case path will draw the "outlines" of the of the grouped lat/lon positions. The layer order and color/size of geom_path() will "overwrite" the white coloured polygon lines.
ggplot(data = counties, aes(x = long, y = lat, group = group)) +
geom_polygon( colour = "white", fill = "lightblue") +
#------------ path layer with inherited "polygon" grouping
geom_path(color = "blue", size = 2)
Next let's create the (zipcodes :) ) aka text labels by averaging the lat/lon values for the different polygon segment points.
cnames <- counties %>%
group_by(subregion) %>%
summarise(long = mean(long), lat = mean(lat) # averaging for "mid"-point
)
> cnames
# A tibble: 100 x 3
subregion long lat
<chr> <dbl> <dbl>
1 alamance -79.4 36.0
2 alexander -81.2 35.9
3 alleghany -81.1 36.5
Now add a geom_text() layer to show the (zip codes) aka subregion names.
ggplot(data = counties, aes(x = long, y = lat, group = group)) +
geom_polygon( colour = "white", fill = "lightblue") +
geom_path(color = "blue", size = 2) +
# --------------- adding a geom_text() layer
geom_text(data = cnames, aes(x = long, y = lat), color = "green")
## ------- uuummmppfff throws an error
Error in FUN(X[[i]], ...) : object 'group' not found
This throws an error. So why is that? Implicitly the group aesthetic in the ggplot() call is understood by geom_polygon() and geom_path() ... however geom_text() has issues with.
Moving the group aesthetics to the polygon layer ...
ggplot(data = counties, aes(x = long, y = lat)) +
geom_polygon( aes(group=group), colour = "white", fill = "lightblue") +
geom_path(color = "blue", size = 2) +
geom_text(data = cnames, aes(x = long, y = lat, label = "subregion"), color = "green")
does the trick but corrupts the geom_path() layer.
What happens here is that the data points (i.e. lat/lon) are no longer grouped, and ggplot connects the end points in the order they appear in the counties data frame. These are the jig-jag lines you see across your plot.
Accordingly, you would need to put another aes(group=group) for the geom_path layer! ... assuming you really want the path for the outlines.
ggplot(data = counties, aes(x = long, y = lat)) +
geom_polygon( aes(group=group), colour = "white", fill = "lightblue") +
geom_path(aes(group=group), color = "blue", size = 2) +
geom_text(data = cnames, aes(x = long, y = lat, label = "subregion"), color = "green")
Obviously the color = "white" is overwritten by the geom_path() call. You may skip one or the other.
As a rule of thumb, ggplot works well with "long" data tables. The moment you add a 2nd (or more other data objects) make sure to track which aesthetics are required and/or inherited from one layer to the other. In your original example, you could move the geom_path() upwards to have the group = group aesthetics from the geom_polygon() call.
In case of doubt, always populate the data = ... and aes() for each layer before combining (and inheriting) parameters across layers.
Good luck!