R language can draw with a fixed number of Australian Geographic Information map
A few days ago I was looking to draw a map, he gave me some number of people on every continent of Australia, the corresponding number of changes every year, I hope this person can do to move up, the data is probably this:
Figure I
Then I made out of this result is the following:
Figure II
the color corresponding data (in fact, this view is wrong, I do it because this is a moving map, so R are superimposed out, the result is like this can keep the earth)
I am writing this blog is to share how I write out, have ideas, I am also a rookie, so some are detours. No wonder, then great God.
The first step must be to find Australia's GIS data
To draw a map, you must get the map data, such biased data, how is it possible to find, map data plotly think the first step in the package, but then checked for a long time, did not find, I think, said on StackOverflow, though plotly American painting map well, but draw other countries Australia, the map will serrated, so I gave up.
Then find ggmap, I'm thinking that this certainly can, and the result was not as ggmap to download google maps, I was in China, how could download, and (mostly do not know how to use Google developer account).
Then find a lot, it seems a lot of sites, only to find, R language has two packages, one oz, is a ozmaps, ha ha ha, finally found, but found the two packages, how do I know how to extract data ? ? These pages are also written a detailed process, then I run site inside the code, finally understood.
The first step in the installation package is loaded package
install.packages(c("devtools", "dplyr", "oz", "ozmaps", "readxl"))
#下面通过library安装的包都要安装
devtools::install_github("thomasp85/transformr")
library(oz)
library(ozmaps)
library(ggplot2)
library(sf)
library(readxl)
library(dplyr)
library(gganimate)
library(tweenr)
Load map data
#地图数据
sf_oz <- ozmap_data("states")
After the above loading is complete, you can see in Rstudio Variables window variables, basically results as shown below:
see variable window on the right, you can see sf_oz data, look at the left side of the window, you can see, there are two variables sf_oz, a nAME is variable, is a variable geometry, I think this means: each name corresponds to a geographic location data.
看图一，图一的表的行名字就是各个州的名字，每一年对应的人数，就是要画出来的数据，这个图1的表的各个州名字和sf-oz变量名字要一一匹配，可以发现，客户给的数据没有sf-oz里面第九行（其他殖民地这个地区，所以，将这删掉。
sf_oz <- sf_oz[-9, ]
删掉之后，sf-oz数据如下：
图三
图一里面的名称都是中文，要翻译成英文然后一一对应，然后将这个客户给的数据和地理信息数据进行合并，得到如下结果：
data_state <- read_xlsx("中国留学生澳洲地区分布.xlsx")
names(data_state) <- c("year", (tibble::as_tibble(sf_oz))[, 1]$NAME)
data_state <- data_state[c(1:11), ]
value = as.matrix(data_state[, -1])
dim(value) <- 88
tidy_data <- tibble(value = value,
year = rep(data_state$year, time = 8),
NAME = rep(unlist(names(data_state)[-1]), each = 11)
)
test_data <- left_join(sf_oz, tidy_data, by = "NAME")
可以看出来，每一个地点，每一年，对应的人数对应的地理位置，都在一行中，这样的数据叫做长数据，我们上面代码的思路就是将客户给的宽数据转换成长数据，然后再将地理信息进行合并。就得到上面的结果了，如果不知道什么叫宽数据和长数据，可以看一下一个人写的[R语言宽数据和长数据如何转换]: https://www.sohu.com/a/110685641_464826: （我当时写的代码不行，太丑了，不知道如何转换宽长数据，导致写的代码很丑，当然现在会了）
上面就是数据处理结果，接下来就是画图：
画图部分
画图部分没有难点，主要就是注意如何将动图进行逐帧播放，如何设置，这个其实也简单，注意一些参数。然后要用到上面的gganimate包，为什么要加载tweenr包，我也不知道，因为如果不加载这个，好像运行有错误，。
anim <- ggplot(data = test_data) geom_sf(aes(fill = value))
theme(plot.title = element_text(hjust = 0.5))
scale_fill_continuous(low = "white", high = "darkblue")
transition_states(year) ggtitle("第{closest_state}年中国留澳学生分布")
animate(anim, 100, fps = 20, width = 1200, height = 1000,
renderer = gifski_renderer("australia.gif")) #改变第一100，可以改变帧数
上面代码就是将一个动态图画出来，怎么画的？？我来一一解答：
ggplot(data = test_data)这个肯定知道，会用ggplot2包画图的都知道要搞这个，毕竟是底层 geom_sf(aes(fill = value))这你买了的geom_sf是表示图层的基础不是普普通通的x,y坐标轴了，而是地理信息图层，这个geom_sf函数是sf包的函数，要注意啦，里面的aes(fill = value)意思说，对不同的地区（区块）用test_data里面的value变量进行填充，这样就显而易见的将我们的人数填充到每个州了，ggtitle("第{closest_state}年中国留澳学生分布") theme(plot.title = element_text(hjust = 0.5))这句话里面的ggtitle是给图加个标题，theme(plot.title = element_text(hjust = 0.5))是说让这个图的标题居中，theme是对ggplot2的一些个性化设置，
plot.title是意味着对标题进行设置，hjust意味着对水平调整，0.5就是居中的意思，1就是右对齐，0就是左对齐。实际上如果test_data只是一年的数据就已经可以了，但是要动起来，必须加上transition_states(year)这句话意思就是说，这个数据通过year这个变量来刷新，每一帧的变化，都是和时间有关。
有些人可能就疑惑了，如果我希望标题显示和年一起刷新，怎么办？？？，（就是当数据显示2017年的时候，标题就显示为“第2017年人数”，当数据刷新到2019年是，标题就是为“第2019年人数”）
实际上，这个gganimate作者就已经考虑到了，细心的朋友已经发现了，ggtitle函数里面有个{closest_state}，哈哈哈，这个就是不断地变化的东西，当你状态是2008， 这个{closest_state}就显示2008， 你改成别的东西，还显示错误（这个我就当时改了，后来才知道，这个变量名字，不是随便改的）。
最后，就可以了。整个代码完成。
图也动起来了。
数据和代码在哪里：
https://github.com/yuanzhoulvpi2017/plot_data
