使用雷达/极坐标图来描绘单个数据点 [英] Using radar/polar graphs to depict individual datapoints
问题描述
这是我以前的帖子：
现在问题开始时，我想要使用这些形状作为数据点。
@brian善意地评论和暗示如何处理这个。
<现在，我们希望将数据留在图形后面并查找正确的数据。
foo< - ggplot_build （g）$ data
＃这是foo的第6个列表。
$：'data.frame'：20 obs。 8个变量：
.. $ color：chr [1:20]＃FF5A5F＃FF5A5F＃FF5A5F＃FF5A5F...
.. $ x：num [1 ：20] 0 0.72 -0.687 0 0 ...
.. $ y：num [1:20] 1.111 -0.416 -0.397 1.111 0.611 ...
.. $ group：atomic [1： 20] 1 1 1 1 2 2 2 2 3 3 ...
.. ..- attr（*，n）= int 5
.. $ PANEL：因子w / 1等级 1：1 1 1 1 1 1 1 1 1 1 ...
.. $ size：num [1:20] 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 ...
.. $ linetype：num [1:20] 1 1 1 1 1 1 1 1 1 1 ...
.. $ alpha：logi [1:20]不适用不适用不适用不适用...
您想将此列表作为数据框提取。我可以发现这个清单的原因是我关注了颜色信息（例如 FF5A5F ）。
mydf < - foo [[6]]
color xy group PANEL size linetype alpha
1＃FF5A5F 0.0000000 1.1111111 1 1 1.5 1 NA
2＃FF5A5F 0.7203077 -0.4158698 1 1 1.5 1 NA
3＃FF5A5F -0.6868360 -0.3965450 1 1 1.5 1 NA
4＃FF5A5F 0.0000000 1.1111111 1 1 1.5 1 NA
5＃ FFB400 0.0000000 0.6111111 2 1 1.5 1 NA
6＃FFB400 0.4286803 -0.2474987 2 1 1.5 1 NA
7＃FFB400 -0.2584135 -0.1491951 2 1 1.5 1 NA
8＃FFB400 0.0000000 0.6111111 2 1 1.5 1 NA
9＃007A87 0.0000000 1.1111111 3 1 1.5 1 NA
10＃007A87 0.7203077 -0.4158698 3 1 1.5 1 NA
11＃007A87 -0.4726248 -0.2728700 3 1 1.5 1 NA
12＃007A87 0.0000000 1.1111111 3 1 1.5 1 NA
13＃8CE071 0.0000000 0.1111111 4 1 1.5 1 NA
14＃8CE071 0.2467912 -0.1424850 4 1 1.5 1 NA
15 ＃8CE071 -0.2278119 -0.1315273 4 1 1.5 1 NA
16＃8CE071 0.0000000 0.1111111 4 1 1.5 1 NA
17＃7B0051 0.0000000 0.1111111 5 1 1.5 1 NA
18＃7B0051 0.2595364 -0.1498434 5 1 1.5 1 NA
19＃7B0051 -0.1268266 -0.0732234 5 1 1.5 1 NA
20＃7B0051 0.0000000 0.1111111 5 1 1.5 1 NA
让我们来确认一下这些数据是否能够再现ggradar图形中的五个三角形。
gg <-ggplot（data = mydf，aes（x = x，y = y，group = group，color = factor（group）））+
geom_path（show.legend = FALSE）+
theme_bw（）
总之，你想破解ggplot数据来提取你想要的数据。此方法允许您保存所需形状的数据。我希望这是你以后的样子。
This is following my previous post: How to measure the area of a polygon in ggplot2?
What I would like now to do is to use the resulting radar charts as data points in a totally separate scatterplot or maybe save the object and use it later on graphs to depict the shape of the actual data.
I have many files and they typically look like this. They range from 1 to many in col number and the fractions are always in descending format. ie #1 always contributes most in any given file.
myfile
01 0.31707317
02 0.12195122
03 0.09756098
04 0.07317073
05 0.07317073
06 0.07317073
07 0.07317073
08 0.07317073
09 0.04878049
10 0.04878049
Here, I would like to plot a radar chart for each of these files individually and just get the shape with no grid around it. ggradar is the closest and the nicest I have found.
ggradar(as.matrix(t(radar)), group.point.size = 2, grid.line.width = 0, grid.max = 1.0, grid.mid = 0,
grid.min = 0, axis.line.colour = "white", axis.label.size = 0, grid.label.size = 0,
centre.y = 0, background.circle.colour = "white", group.colours = "black", group.line.width = 1)
Now the problem starts when I want to use these shapes as data points.
@brian kindly commented and hinted as to how to approach this.
R::ggplot2::geom_points: how to swap points with pie charts?
Small ggplot2 plots placed on coordinates on a ggmap
Inspired by these comments, I can now save plots as one column along with the filename and keep adding the new files using the same pipeline.
for (file in file_list){
#eliminate the empty files (they contain only the header)
if (file.size(file) > 420){
# if the merged dataset does exist, append to it
if (exists("dfradar")){
radarfile <-read.table(file, header=TRUE, sep="\t")
radarfile1 <- as.data.frame(as.numeric(radarfile[,3]))
rownames(radarfile1) <- c(1:nrow(radarfile))
dfradar1 <- ggradar(t(radarfile1), group.point.size = 1, grid.line.width = 0, grid.max = 1, grid.mid = 0,
grid.min = 0, axis.line.colour = "white", axis.label.size = 0, grid.label.size = 0,
centre.y = 0, background.circle.colour = "white", group.colours = "black", group.line.width = 0.5) +
theme(legend.position = "none")
dfradar1 <- cbind(substring(file,11), dfradar1)
dfradar <- rbind(dfradar, dfradar1)
}
# if the merged dataset doesn't exist, create it
if (!exists("dfradar")){
radarfile <- read.table(file, header=TRUE, sep="\t")
radarfile1 <- as.data.frame(as.numeric(radarfile[,3]))
rownames(radarfile1) <- c(1:nrow(radarfile))
dfradar <- ggradar(t(radarfile1), group.point.size = 1, grid.line.width = 0, grid.max = 1, grid.mid = 0,
grid.min = 0, axis.line.colour = "white", axis.label.size = 0, grid.label.size = 0,
centre.y = 0, background.circle.colour = "white", group.colours = "black", group.line.width = 0.5) +
theme(legend.position = "none")
dfradar <- cbind(substring(file,11), dfradar)
rm(radarfile)
rm(radarfile1)
}
} }
Now when I want to save the grobs, I can't; I get:
"Error in mutate_impl(.data, dots) : Evaluation error: attempt to apply non-function."
or:
"Error in plot_clone(plot) : attempt to apply non-function"
# merge the df with another df containing all other
# variables that I wanna use in my scatterplot
dfradar_merge <- merge(dfradar, Cases, all=FALSE)
dfradar_merge <- dfradar_merge %>% mutate(radargrobs = list(annotation_custom(ggplotGrob(radarplots)),
xmin = as.numeric(Gender), xmax = as.numeric(Gender)*1.2,
ymin = as.numeric(Age) , ymax = as.numeric(Age)*1.2)))
I think if I can get this part straight and pass the variables as xmin and ymin, I should be able to do what I need...
Any ideas are really appreciated. It has been a headache for me.
I came up with the following idea. I used the ggradar package to create a graphic. Then, I looked into the data frames staying behind the graphic with ggplot_build(). It seems that there are 13 lists behind the graphic. I needed to investigate which list is the right one containing polygon data. I found that the 6th list is the one you want to extract and save for your future use. Let me show you what I have done.
library(dplyr)
library(ggradar)
library(scales)
# I modified the code from https://github.com/ricardo-bion/ggradar to get a graphic.
mtcars %>%
mutate_all(rescale) %>%
mutate(group = rownames(mtcars)) %>%
slice(5:9) %>%
select(1:4) -> mtcars_radar
g <- ggradar(mtcars_radar)
The code above generates the following graphic. There is something wrong with my computer or something else, and I do not see a complete graphic. But this is not an issue for this question.
Now, we want to get the data staying behind the graphic and look for the right data.
foo <- ggplot_build(g)$data
# This is the 6th list in foo.
$ :'data.frame': 20 obs. of 8 variables:
..$ colour : chr [1:20] "#FF5A5F" "#FF5A5F" "#FF5A5F" "#FF5A5F" ...
..$ x : num [1:20] 0 0.72 -0.687 0 0 ...
..$ y : num [1:20] 1.111 -0.416 -0.397 1.111 0.611 ...
..$ group : atomic [1:20] 1 1 1 1 2 2 2 2 3 3 ...
.. ..- attr(*, "n")= int 5
..$ PANEL : Factor w/ 1 level "1": 1 1 1 1 1 1 1 1 1 1 ...
..$ size : num [1:20] 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 ...
..$ linetype: num [1:20] 1 1 1 1 1 1 1 1 1 1 ...
..$ alpha : logi [1:20] NA NA NA NA NA NA ...
You want to extract this list as a data frame. The reason why I could spot this list is that I was paying attention to color information (e.g., FF5A5F).
mydf <- foo[[6]]
colour x y group PANEL size linetype alpha
1 #FF5A5F 0.0000000 1.1111111 1 1 1.5 1 NA
2 #FF5A5F 0.7203077 -0.4158698 1 1 1.5 1 NA
3 #FF5A5F -0.6868360 -0.3965450 1 1 1.5 1 NA
4 #FF5A5F 0.0000000 1.1111111 1 1 1.5 1 NA
5 #FFB400 0.0000000 0.6111111 2 1 1.5 1 NA
6 #FFB400 0.4286803 -0.2474987 2 1 1.5 1 NA
7 #FFB400 -0.2584135 -0.1491951 2 1 1.5 1 NA
8 #FFB400 0.0000000 0.6111111 2 1 1.5 1 NA
9 #007A87 0.0000000 1.1111111 3 1 1.5 1 NA
10 #007A87 0.7203077 -0.4158698 3 1 1.5 1 NA
11 #007A87 -0.4726248 -0.2728700 3 1 1.5 1 NA
12 #007A87 0.0000000 1.1111111 3 1 1.5 1 NA
13 #8CE071 0.0000000 0.1111111 4 1 1.5 1 NA
14 #8CE071 0.2467912 -0.1424850 4 1 1.5 1 NA
15 #8CE071 -0.2278119 -0.1315273 4 1 1.5 1 NA
16 #8CE071 0.0000000 0.1111111 4 1 1.5 1 NA
17 #7B0051 0.0000000 0.1111111 5 1 1.5 1 NA
18 #7B0051 0.2595364 -0.1498434 5 1 1.5 1 NA
19 #7B0051 -0.1268266 -0.0732234 5 1 1.5 1 NA
20 #7B0051 0.0000000 0.1111111 5 1 1.5 1 NA
Let's confirm if this data reproduces the five triangles in the ggradar graphic.
gg <- ggplot(data = mydf, aes(x = x, y = y, group = group, color = factor(group))) +
geom_path(show.legend = FALSE) +
theme_bw()
In summary, you want to hack ggplot data to extract the data you want. This method allows you to save data for the "shape" that you want. I hope this is what you are after.
这篇关于使用雷达/极坐标图来描绘单个数据点的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
