Compare cities by area in ggplot2 on map

Question

I need compare ukrainian cities by area in ggplot2. But all of they plot in the same size.

Can you help me?

I use this code:

library(ggplot2)

cities <- read.csv("https://raw.githubusercontent.com/savchukidze/maps/master/cities.csv", stringsAsFactors = F)


png("compare_cities.png", height = 3000, width = 2500)

ggplot()+

   geom_polygon(data = cities, aes(long, lat, group = group), 
                color = "black", size = 2.5, fill = "grey", alpha = 1)+
   facet_wrap(~city, scales = "free")

dev.off()

Athanasia Mowinckel · Answer 1 · 2018-12-19T11:48:14.083

1

This is a little tricky, because you need each panel to have "free" space, but cannot combine that with coord_fixed.

You'll need to place them all on the scale for coordinates.

library(tidyverse)
cities %>% 
    group_by(city) %>% 
    mutate(long=long-mean(long), lat=lat-mean(lat)) %>% 
    ggplot(aes(long, lat, group=city)) + 
        geom_polygon() + 
        facet_wrap(~city)

edited Dec 19 '18 at 11:48

answered Dec 19 '18 at 11:08

Athanasia Mowinckel

432
2
8

In the second version, all of them are in the same size, unfortunately( – Augusto Chile Dec 19 '18 at 11:43
true, I'll remove it. But the first option they are not the same size. – Athanasia Mowinckel Dec 19 '18 at 11:46
yep, thanks, your code is simple to understand! I think i'll use Photoshop to put all cities in the center of plots. – Augusto Chile Dec 19 '18 at 11:58

score 1 · Accepted Answer · answered Dec 19 '18 at 11:34

It is a bit messy, but you could do something like this to plot invisible points at the maximum range for each city, to keep the scales the same...

library(tidyverse)
ranges <- cities %>% 
  group_by(city) %>% 
  summarise(minlat=min(lat),              #get limits for each city
            maxlat=max(lat),
            minlon=min(long),
            maxlon=max(long)) %>% 
  mutate(rangelat=maxlat-minlat,          #calculate range required for each
         rangelon=maxlon-minlon,
         centrelat=(maxlat+minlat)/2,     #calculate centre point of plot
         centrelon=(maxlon+minlon)/2) %>% 
  ungroup() %>% 
  mutate(bottom=centrelat-max(rangelat)/2,#calculate box size based on max range
         top=centrelat+max(rangelat)/2,
         left=centrelon-max(rangelon)/2,
         right=centrelon+max(rangelon)/2)

ggplot()+
  geom_polygon(data = cities, aes(long, lat, group = group), 
               color = "black", size = 2.5, fill = "grey", alpha = 1)+
  geom_point(data=ranges, aes(x=left,y=bottom), alpha=0)+ #invisible bottom left point
  geom_point(data=ranges, aes(x=right,y=top),alpha=0)+    #invisible top right point
  facet_wrap(~city,scales = "free")

It's amazing. Messy in not a problem. You are genius! So, one more question. May be you know, can I calculate area of this cities by coordinates and arrange cities by size of area? — Augusto Chile, Dec 19 '18 at 11:51
I'm sure it can be done, but I'm not sure how without a bit of further research. It might be worth asking it as a separate question on here. — Andrew Gustar, Dec 19 '18 at 14:12

Compare cities by area in ggplot2 on map

2 Answers2