R中的地热图

Posted 2021-04-15

我想在R中创建美国地图，并根据度量以热图类型的方式对州进行颜色编码。我知道如何使用googleVis api来执行此操作，但是我无法使用代码，并且没有过渡也不是那么好。完成这项工作的最快方法是什么？我对地图包很熟悉，但是我无法获得合作的颜色。我相信这被称为choropleth地图。

答案

ggplot2软件包中有完整的示例，请参见?map_data。

library(ggplot2)
example(map_data)

另一答案

（希望答案可能仍然对某人有用）

RevolutionAnalytics使用example功能具有出色的地图可视化spplot()。这是那里的图片：

<< img src =“ https://image.soinside.com/eyJ1cmwiOiAiaHR0cHM6Ly9pLnN0YWNrLmltZ3VyLmNvbS9QWmw1ai5wbmcifQ==” alt =“在此处输入图像描述”>

另一答案

在UScensus2000tract软件包中有一个以人口为导向的螺旋形的例子。

另一答案

代码

# Transform dataset to correct format
crimes <- data.frame(state = tolower(row.names(USArrests)), USArrests)
crimes

#  Map data   
#  install.packages("maps") remember to install these two packages if you  
#  install.packages("mapproj") do not have them yet

library(mapproj)
library(maps)

states_map <- map_data("state")
states_map

# Merge datasets together 
crime_map <- merge(states_map, crimes, by.x = "region", by.y = "state")

# After merging, the order has changed, which leads to polygons drawn 
# in the incorrect order. Let's sort it 
crime_map

library(dplyr) # for arrange() function 

# Sort by group, then order 
crime_map <- arrange(crime_map, group, order)
crime_map

# Now data can be plotted
library(ggplot2)

plot1 <- ggplot(crime_map, aes(x = long, y = lat, group = group, fill = Assault)) + 
                    geom_polygon(colour = "black") + 
                    coord_map("polyconic")

plot1

# Add title 
plot1 <- plot1 + 
                ggtitle("                       Proportion of crimes in the USA")

plot1

# Improve on colours 
plot1 <- plot1 + 
    scale_fill_gradient2(low = "#559999", mid = "grey", high = "#BB650B",
                                                    midpoint = median(crimes$Assault))
plot1

# If you want white transparent backgroud only
plot1 <- plot1 + 
         theme_void()

plot1

# Note: if RStudio gives you this error when ploducing plot, then use this and try 
# again
devAskNewPage(ask = FALSE)

# Special thanks to Winston Chang and other ggplot developers at RStudio who made made 
#  many of these codes