2

I am trying to make a grid containing maps of megaregions in the us. I create a SpatialPolygonDataframe from a shape file. then convert it into a data.frame to use ggplot2. as soon as I add the data into the frame, the polygon plots. the file containing SpatialPolygon and the data frame are here: https://drive.google.com/open?id=1kGPZ3CENJbHva0s558vWU24-erbqWUGo the code is as follow:

load("./data.rda")
prop.test <- proptest.result[which(proptest.result$variable=="Upward N"),]

#transforming the data
# add to data a new column termed "id" composed of the rownames of data
shape@data$id <- rownames(shape@data)
#add data to our 
shape@data <- data.frame(merge(x = shape@data, y = prop.test, by.x='Name', by.y="megaregion"))

# create a data.frame from our spatial object
mega.prop <- fortify(shape)
#merge the "fortified" data with the data from our spatial object
mega.prop.test <- merge(mega.prop, shape@data, by="id")

Plotting the first one (mega.prop) works fine:

ggplot(data = mega.prop, aes(x=long, y=lat, group=group), fill="blue")+
    geom_polygon()

the plot before adding information dataframe

but plotting after adding the analytics data:

ggplot(data = mega.prop.test, aes(x=long, y=lat, group=group), fill="blue")+
    geom_polygon()

The plot after adding analytical data

In the new plot:

  • The filling of polygons is messed up. (Is it about the order of points?how?)
  • two of the polygons are totally missed.

What is the problem? Thank you very much for your help.

BobbyF
  • 431
  • 1
  • 7
  • 19
  • 1
    the base version of `merge` messes up point order. Use `data.table::merge` or `dplyr::left_join` instead. (Or add an "order" column first and sort by it afterwards.) – Gregor Thomas Jan 26 '18 at 22:10

1 Answers1

3

Use geom_map() (which requires a slight tweak of your shapefile for some reason) so you don't have to do the merge/left join.

Also, you merged a great deal of different factors, not sure which ones you want to plot.

Finally, it's unlikely the coastal areas need that fine level of detail. rgeos::gSimplify() will definitely speed things up and you're already distorting areas, so a smaller bit of additional distortion won't impact the results.

library(ggplot2)
library(tidyverse)

shape_map <- tbl_df(fortify(shape, region="Name"))
colnames(shape_map) <- c("long", "lat", "order", "hole", "piece", "region", "group")

prop.test <- proptest.result[which(proptest.result$variable=="Upward N"),]

ggplot() +
  geom_map(data=shape_map, map=shape_map, aes(long, lat, map_id=region)) +
  geom_map(
    data=filter(prop.test, season=="DJF"),
    map=shape_map, aes(fill=prop.mega, map_id=megaregion)
  )

enter image description here

hrbrmstr
  • 77,368
  • 11
  • 139
  • 205
  • Thank you very much. It was a great help, and solved this problem completely. I still have one problem in the follow, that I just put it into [this question;](https://stackoverflow.com/q/48472163/1813491). I will be very grateful if you can take a look at that, too. Thank you so much, agian. – BobbyF Jan 27 '18 at 03:03