An R package storing the geospatial shapefile (vector data) of China administrative divisions to the county/district-level.
To install CRAN version 安装CRAN正式版本:
install.packages("mapchina")To install the most updated version of the 'dev' branch 安装开发中的最新版本:
if (!require(devtools)) {
install.packages("devtools")
}
devtools::install_github("xmc811/mapchina", ref = "dev")library(mapchina)
head(china)# Output 输出
Simple feature collection with 6 features and 13 fields
geometry type: MULTIPOLYGON
dimension: XY
bbox: xmin: 115.4248 ymin: 39.44473 xmax: 116.8805 ymax: 41.05936
geographic CRS: WGS 84
# A tibble: 6 x 14
Code_County Code_Perfecture Code_Province Name_Province Name_Perfecture Name_County Pinyin Pop_2000 Pop_2010 Pop_2017 Pop_2018 Area Density
<chr> <chr> <chr> <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 110101 1101 11 北京市 NA 东城区 Dōngc… 881763 919253 NA 822000 41.8 19670.
2 110102 1101 11 北京市 NA 西城区 Xīché… 1232823 1243315 NA 1179000 50.5 23360.
3 110114 1101 11 北京市 NA 昌平区 Chāng… 614821 1660501 NA 2108000 1342 1571.
4 110115 1101 11 北京市 NA 大兴区 Dàxīn… 671444 1365112 NA 1796000 1053 1706.
5 110111 1101 11 北京市 NA 房山区 Fángs… 814367 944832 NA 1188000 1995 595.
6 110116 1101 11 北京市 NA 怀柔区 Huáir… 296002 372887 NA 414000 2123 195.
# … with 1 more variable: geometry <MULTIPOLYGON [°]>The main data object china is a dataframe, with each row as one county/district level administrative division of China.
To plot the map of a particular region, you can filter() codes or names to subset the dataframe.
The codes are stored in variables Code_Province, Code_Perfecture, and Code_County,
and the names are stored in Name_Province, Name_Perfecture, and Name_County.
The codes and names follow the 3-level hierarchy: Province (2-digit), Perfecture (4-digit), and County (6-digit).
在矢量数据china中，行政区划的中文名与代码均分为三级：省级、地级与县级，可使用filter()进行任意筛选。
注意：直辖市、特别行政区、台湾地区、省直管市以及其他特殊行政区域的地级区域名称Name_Perfecture为NA。用户可根据作图需要进行修改。
library(sf)
st_write(china, "your/path/to/china.shp", layer_options = "ENCODING=UTF-8")Since the shapefile data is also a dataframe, it can be plotted by ggplot grammer of graphics. The geometric object is geom_sf().
library(tidyverse)
library(sf)
df <- china %>%
filter(Code_Province %in% c("11","12","13"))
ggplot(data = df) +
geom_sf(aes(fill = rank(Density))) +
scale_fill_distiller(palette = "BuPu", direction = 1) +
theme_bw() +
theme(legend.position = "none")
New data can be added to the shapefile dataframe as new variables
df$Var <- runif(nrow(df))
ggplot(data = df) +
geom_sf(aes(fill = Var)) +
scale_fill_distiller(palette = "YlOrRd") +
theme_bw() +
theme(legend.position = "none")
We use greedy coloring algorithm to solve the problem. The function generate_map_colors() takes a shapefile dataframe as input and outputs a list of index for filling colors.
df2 <- china %>%
filter(Code_Province %in% c("32"))
ggplot(data = df2) +
geom_sf(aes(fill = factor(generate_map_colors(df2)))) +
scale_fill_brewer(palette = "Set3") +
theme_bw() +
theme(legend.position = "none")
The geometry of county-level divisions can be merged to higher level divisions by functions group_by(), summarise(), and sf::st_union().
县级行政区可被合并为地级或省级行政区进行作图。
df3 <- china %>%
filter(Code_Province %in% as.character(31:36))
df3 <- df3 %>%
group_by(Name_Province) %>%
summarise(geometry = st_union(geometry))
ggplot(data = df3) +
geom_sf(aes(fill = Name_Province)) +
scale_fill_brewer(palette = "Set3") +
theme_bw() +
theme(legend.position = "none")
To add Chinese characters for the map, R package showtext is required.
if (!require("showtext")) {
install.packages("showtext")
showtext::showtext_auto()
ggplot(data = df3) +
geom_sf(aes(fill = Name_Province)) +
scale_fill_brewer(palette = "Set3") +
theme_bw() +
theme(legend.position = "none") +
geom_sf_label(aes(label = Name_Province))
showtext::showtext_auto()
df4 <- china %>%
filter(Code_Province %in% c("32","34"))
df4_prov <- df4 %>%
group_by(Name_Province) %>%
summarise(geometry = st_union(geometry))
df4_perf <- df4 %>%
group_by(Name_Perfecture) %>%
summarise(geometry = st_union(geometry))
ggplot() +
geom_sf(data = df4_perf,
aes(fill = factor(generate_map_colors(df4_perf))),
linetype = "solid", size = 0.5) +
scale_fill_brewer(palette = "Pastel1") +
geom_sf(data = df4, alpha = 0, linetype = "dashed", size = 0.2) +
geom_sf(data = df4_prov, alpha = 0, linetype = "solid", size = 1.2) +
geom_sf_label(data = df4_perf, aes(label = Name_Perfecture)) +
theme_bw() +
theme(legend.position = "none")
