Graphing lat/long data points on US map 50 states including Alaska & Hawaii

Question

I'm trying to create a bubble plot similar to this bubble plot from plot.ly with all 50 US states, with Alaska and Hawaii relocated.

The data set would have 5 variables, zip code, latitude longitude (merged on from the zipcode package), amount and score. The amount and score variables would be represented by the size and color of the bubble.

I've been trying to do this using ggplot. The first method uses the maps package. However, I don't know how I would move the Alaska and Hawaii and refit it, as well as draw the border of the US.

This is what I get.

US <- map_data("world") %>% filter(region=="USA")


p <- ggplot() + 
  geom_polygon(data = US, aes(x=long, y = lat, group = group), fill="grey", alpha=0.3) +
  geom_point(data = df, alpha = 0.5, aes(x=df$lon, y=df$lat, size=df$manual_prm, color=df$wcng_score)) +
  scale_size_continuous(name="Premium", trans="log",  range=c(0.1,4)) +
  scale_alpha_continuous(range=c(0.1, 0.9)) +
  scale_color_viridis(name = "Scores", trans="log", option="magma") +
  theme_void() + 
  coord_map() +
  ggtitle("Scores") +
  theme(
    legend.position = c(0.85, 0.8),
    text = element_text(color = "#22211d"),
    plot.background = element_rect(fill = "#f5f5f2", color = NA), 
    panel.background = element_rect(fill = "#f5f5f2", color = NA), 
    legend.background = element_rect(fill = "#f5f5f2", color = NA),
    plot.title = element_text(size= 16, hjust=0.1, color = "#4e4d47", margin = margin(b = -0.1, t = 0.4, l = 2, unit = "cm"))
  )

The other method involves using albersusa package. However, I get almost the same issue. This version also made the boundaries invisible.

Here is the result.

I left out the scales.

library(tidyverse)
library(albersusa)
library(broom)


USA <- usa_composite(proj="laea")  # creates map projection 
USA_MAP <- tidy(USA, region="name")

q <- ggplot() + 
  geom_map(data = USA_MAP, map = USA_MAP, aes(map_id = id), fill=NA, size = 0.1)+
  geom_point(data = df, alpha = 0.5, aes(x=df$lon, y=df$lat, size=df$manual_prm, color=df$wcng_score)) +
  theme_void() +
  coord_map()

How can I get something similar to what plotly produces?

sebdalgarno · Answer 1 · 2018-03-28T04:49:58.367

I'll follow up on neilfws suggestion to use fiftystater. Unfortunately I don't have your df object so cannot recommend how best to match size, alpha and colours. However, the plot.ly map looks to be in USA Albers equal-area projection so I have also provided code to get the projection looking better. This requires transforming the coordinates with sf and using coord_sf instead of coord_map in ggplot.

library(fiftystater)
library(sf)

epsg <- 102003
US <- st_as_sf(fifty_states, coords =c("long", "lat"), crs = 4326) %>%
  st_transform(epsg) %>%
  cbind(st_coordinates(.))

p <- ggplot() + 
  geom_polygon(data = US, aes(x=X, y = Y, group = group), fill="grey", alpha=0.3) +
  geom_point(data = df, alpha = 0.5, aes(x=df$lon, y=df$lat, size=df$manual_prm, color=df$wcng_score)) +
  scale_size_continuous(name="Premium", trans="log",  range=c(0.1,4)) +
  scale_alpha_continuous(range=c(0.1, 0.9)) +
  scale_color_viridis(name = "Scores", trans="log", option="magma") +
  theme_void() +
  coord_sf(crs = epsg, datum = NA) +
  ggtitle("Scores") +
  theme(
    legend.position = c(0.85, 0.8),
    text = element_text(color = "#22211d"),
    plot.background = element_rect(fill = "#f5f5f2", color = NA), 
    panel.background = element_rect(fill = "#f5f5f2", color = NA), 
    legend.background = element_rect(fill = "#f5f5f2", color = NA),
    plot.title = element_text(size= 16, hjust=0.1, color = "#4e4d47", margin = margin(b = -0.1, t = 0.4, l = 2, unit = "cm"))
  )

Hmm, this is weird but when I try to run this I keep getting the error could not fund function coord_sf. both packages sf and ggplot2 are loaded. — chaose, Mar 28 '18 at 16:45
sorry I should have mentioned that you need the dev version of ggplot2. Install by doing...devtools::install_github("tidyverse/ggplot2"). Note you also need to install 'devtools' package to run 'install_github' — sebdalgarno, Mar 28 '18 at 22:44
On closer inspection, looks like `fiftystater` is more suited to choropleth maps. May be possible to get points on it, but probably more work than I anticipated initially. — neilfws, Mar 29 '18 at 02:29
fiftystater simply provides the polygon layer within the ggplot. Any number of other layers (points, lines, text, etc) can be added using ggplot. It is exactly the same amount of work as if you had some other source for U.S state polygons — sebdalgarno, Mar 29 '18 at 04:11

score 1 · Answer 2 · answered Mar 28 '18 at 18:15

I was able to solve this issue now. The problem was that my coordibates for Alaska and Hawaii were in the original position. What I did was to convert my data to a spatial data points and use points_elided from the albersusa package to convert my coordinates to the correct position with the moved Alaska and Hawaii maps.

I added the following code

library(albersusa)
library(sp)
library(raster)
library(ggplot2)
library(bloom)

# set up reference coordinate system and convert to spatial points
coordinates(df) <- ~lon + lat
latlong = "init=epsg:4326"
proj4string(df) = CRS(latlong)
df <- as.data.frame(points_elided(df))


USA <- usa_composite(proj="laea")  # creates map projection 
USA_MAP <- tidy(USA, region="name")

q <- ggplot() + 
  geom_map(data = USA_MAP, map = USA_MAP, aes(map_id = id), fill=NA, size = 0.1)+
  geom_point(data = df, alpha = 0.5, aes(x=df$lon, y=df$lat, size=df$manual_prm, color=df$wcng_score)) +
  theme_void() +
  coord_map()

I think you need `library(broom)` not `library(bloom)` – dca Jul 08 '22 at 20:30 — dca, Jul 08 '22 at 20:30
And `latlong = "+init=epsg:4326"` with the added + – dca Jul 08 '22 at 20:32 — dca, Jul 08 '22 at 20:32

neilfws · Answer 3 · 2023-08-31T22:43:49.893

-1

You could try installing the fiftystater package, as described here.

UPDATE: I note that fiftystater is no longer available at CRAN and the Github code was last updated in 2016 so my answer is now unlikely to be a good solution.

edited Aug 31 '23 at 22:43

answered Mar 27 '18 at 22:40

neilfws

32,751
5
50
63

2

is this an answer or a comment? – SymbolixAU Mar 27 '18 at 23:42
The problem with the fiftystater is that I believe it is on a state level, not at a latitude longitude coordinate level. I'm looking to map things out using latitude and longitude – chaose Mar 28 '18 at 16:57
sorry i was stupid, i can just use geom_point to plot the points – chaose Mar 28 '18 at 17:10
I used this and the problem is that the Alaska and Hawaii points are still off to the side instead of being shown below the US maps, – chaose Mar 28 '18 at 17:15
neilfws, this answer is now obsolete since the link is no longer available. Please consider removing it or updating it with relevant information instead of a link. Now that I have you here, it'd be appreciated if you pop into https://chat.stackoverflow.com/rooms/252064/r-collective-planning to at least get acquainted with the other RMs. (p.s. I came across this organically when searching for a dupe-target for this question: https://stackoverflow.com/questions/77019202/plotting-points-on-us-map-in-r-by-zip-code-with-alaska-and-hawaii-inset) – M-- Aug 31 '23 at 21:01

Graphing lat/long data points on US map 50 states including Alaska & Hawaii

3 Answers3