如何绘制不同的索引，在 R 中保持一个固定
How to plot different indices keeping one fixed in R
我有一个包含房价指数（销售和租金）的数据框，其中包含 3 个不同版本的 HPI 和中位数价格指数。因此，我有一个带有索引值的变量“值”，“索引”具有 HPI1、HPI2、HPI3 或中值，用于索引类型、销售或出租的“操作”和月份。看起来像这样，但时间更长。
Index
Value
Operation
Month
HPI1
0.9
Sale
01/2020
HPI1
1.1
Rent
02/2020
HPI2
0.89
Sale
01/2020
HPI2
1.12
Rent
02/2020
HPI3
0.85
Sale
01/2020
HPI3
1.22
Rent
02/2020
Median
0.91
Sale
01/2020
Median
1.02
Rent
02/2020
我需要通过运算用中位数绘制每个 HPI。这意味着我需要在一个图中绘制三个图，第一个是 HPI1 + Median，第二个是 HPI2 + Median，第三个是 HPI3 + Median。我在循环中使用 ggplot，但我只设法用四行绘制了 1 个图。我不知道如何通过 HPI 重复情节但保持中位数。有什么想法吗？
非常感谢！
我目前正在使用这段代码（RES是数据框）
for(z in operation
ggplot(RES[operation==z,],aes(x=interaction(month,year,sep = " "),y=Value,group=index),alpha=0.02) + geom_line(aes(linetype=index)) + geom_point() + scale_linetype_manual(values=c("dotted", "solid")) + theme(legend.position='bottom',legend.title=element_blank(),axis.text.x = element_text(angle = 90))+ geom_hline(aes(yintercept=1,colour="red"),show.legend = F)+facet_wrap(~operacion,scales = "free",ncol = 2)
}
这是一个可能的解决方案。
它在许多 HPI 的情况下是可扩展的。
它完全基于 tidyverse。
这个想法是通过使用来自 tidyr 的两个 pivot 命令在每个 HPIn 旁边设置 Median。
您可以使用 facet_grid 或 facet_wrap.
在一张图像中获得多个绘图
解决方案
library(dplyr)
library(tidyr)
library(ggplot2)
df %>%
# transform in date
mutate(Month = as.Date(paste0("01/", Month), format = "%d/%m/%Y")) %>%
# reshape data
pivot_wider(names_from = Index, values_from = Value) %>%
pivot_longer(starts_with("HPI"), names_to = "Index", values_to = "Value") %>%
# plot by HPI
ggplot(aes(x = Month)) +
geom_line(aes(y = Value, colour = Index)) +
geom_line(aes(y = Median, colour = "Median")) +
geom_point(aes(y = Value, colour = Index)) +
geom_point(aes(y = Median, colour = "Median")) +
scale_x_date(date_labels = "%m %Y", date_breaks = "1 month") +
facet_grid(Index~Operation)
图例是多余的。如果您不想要它：删除第二个 geom_line 中的 color = "Median" 并在第一个 geom_line 中添加 show.legend = FALSE。
或者你可以在最后添加+ theme(legend.position = "none")。
数据
# (I just tripled your data)
df <- tibble::tribble(
~Index, ~Value, ~Operation, ~Month,
"HPI1", 0.9, "Sale", "01/2020",
"HPI1", 1.1, "Rent", "02/2020",
"HPI2", 0.89, "Sale", "01/2020",
"HPI2", 1.12, "Rent", "02/2020",
"HPI3", 0.85, "Sale", "01/2020",
"HPI3", 1.22, "Rent", "02/2020",
"Median", 0.91, "Sale", "01/2020",
"Median", 1.02, "Rent", "02/2020",
"HPI1", 0.9, "Sale", "02/2020",
"HPI1", 1.1, "Rent", "03/2020",
"HPI2", 0.89, "Sale", "02/2020",
"HPI2", 1.12, "Rent", "03/2020",
"HPI3", 0.85, "Sale", "02/2020",
"HPI3", 1.22, "Rent", "03/2020",
"Median", 0.91, "Sale", "02/2020",
"Median", 1.02, "Rent", "03/2020",
"HPI1", 0.9, "Sale", "03/2020",
"HPI1", 1.1, "Rent", "04/2020",
"HPI2", 0.89, "Sale", "03/2020",
"HPI2", 1.12, "Rent", "04/2020",
"HPI3", 0.85, "Sale", "03/2020",
"HPI3", 1.22, "Rent", "04/2020",
"Median", 0.91, "Sale", "03/2020",
"Median", 1.02, "Rent", "04/2020")
另一种使用 foreach 循环的方法
注意：我使用了@Edo answer
中提供的示例数据
初始库加载并准备图例供以后使用
library(ggplot2) # For obvious reason - graph
library(dplyr) # For manipulate data a bit
library(foreach) # Foreach Loop
library(cowplot) # For manipulate graphs
# Define the line type Solid for variables and dotted line for Median
line_type <- c("solid", "1111")
names(line_type) <- c("Index (HPI1, HPI2, HPI3)", "Median")
# Generate dummy plot just to extract legend
plot_for_legend <- ggplot(data = RES %>%
mutate(Index = if_else(Index == "Median", "Median",
# If you have more Index than HPI1, HPI2, HPI3, can just simply remove
# the text in parentheses and leave it as Index - make sure you change
# the line_type variable accordingly and adjust all the places needed.
"Index (HPI1, HPI2, HPI3)")),
aes(x = Month, y = Value, group = Index)) +
geom_line(aes(linetype = Index)) +
scale_linetype_manual(values = line_type) +
theme(legend.position = 'bottom', legend.title = element_blank(),
axis.text.x = element_text(angle = 90))
# Extracting legend using get_legend from cowplot package
legend_area <- get_legend(plot_for_legend)
使用 foreach 循环遍历每个操作并绘制它们
operation <- unique(RES[["Operation"]])
operation_plot <- foreach(z = operation) %do% {
# filter data to current operation Rent/ Sale
operation_data <- RES %>% filter(Operation == z)
median_data <- operation_data %>% filter(Index == "Median")
median_line_point <- geom_line(data = median_data,
aes(x = Month, y = Value, group = Index, linetype = "Median"),
alpha=1)
median_point <- geom_point(data = median_data,
aes(x = Month, y = Value))
list_index <- unique(operation_data[["Index"]])
list_index <- list_index[list_index != "Median"]
list_plot <- foreach(i_index = list_index) %do% {
ggplot(data = operation_data %>% filter(Index == i_index),
aes(x = Month, y = Value, group = Index,
# in case you change the line_type make sure to change the name
# here accordingly
linetype ="Index (HPI1, HPI2, HPI3)"),
alpha=1) +
geom_line(linetype = "solid") +
geom_point() +
median_line_point + median_point +
scale_linetype_manual(values = line_type) +
geom_hline(aes(yintercept = 1,colour = "red"), show.legend = F) +
ylab(i_index) +
theme(legend.position = "none")
}
# put all the plot in the list into one
current_plot <- plot_grid(plotlist = list_plot, ncol = 2)
# Generate title for operation
title <- ggdraw() +
draw_label(
sprintf("%s operation", z),
fontface = 'bold',
x = 0,
hjust = 0
) +
theme(
# add margin on the left of the drawing canvas,
# so title is aligned with left edge of first plot
plot.margin = margin(0, 0, 0, 7)
)
# Combine title, main plot & legend into one
current_plot_w_legend <- plot_grid(title, current_plot, legend_area, nrow = 3, rel_heights = c(.5, 10, 1))
}
# Set names of the plot variables for easier access
names(operation_plot) <- operation
这里有一些示例数据是输出
我有一个包含房价指数（销售和租金）的数据框，其中包含 3 个不同版本的 HPI 和中位数价格指数。因此，我有一个带有索引值的变量“值”，“索引”具有 HPI1、HPI2、HPI3 或中值，用于索引类型、销售或出租的“操作”和月份。看起来像这样，但时间更长。
| Index | Value | Operation | Month |
|---|---|---|---|
| HPI1 | 0.9 | Sale | 01/2020 |
| HPI1 | 1.1 | Rent | 02/2020 |
| HPI2 | 0.89 | Sale | 01/2020 |
| HPI2 | 1.12 | Rent | 02/2020 |
| HPI3 | 0.85 | Sale | 01/2020 |
| HPI3 | 1.22 | Rent | 02/2020 |
| Median | 0.91 | Sale | 01/2020 |
| Median | 1.02 | Rent | 02/2020 |
我需要通过运算用中位数绘制每个 HPI。这意味着我需要在一个图中绘制三个图，第一个是 HPI1 + Median，第二个是 HPI2 + Median，第三个是 HPI3 + Median。我在循环中使用 ggplot，但我只设法用四行绘制了 1 个图。我不知道如何通过 HPI 重复情节但保持中位数。有什么想法吗？
非常感谢！
我目前正在使用这段代码（RES是数据框）
for(z in operation
ggplot(RES[operation==z,],aes(x=interaction(month,year,sep = " "),y=Value,group=index),alpha=0.02) + geom_line(aes(linetype=index)) + geom_point() + scale_linetype_manual(values=c("dotted", "solid")) + theme(legend.position='bottom',legend.title=element_blank(),axis.text.x = element_text(angle = 90))+ geom_hline(aes(yintercept=1,colour="red"),show.legend = F)+facet_wrap(~operacion,scales = "free",ncol = 2)
}
这是一个可能的解决方案。
它在许多 HPI 的情况下是可扩展的。
它完全基于 tidyverse。
这个想法是通过使用来自 tidyr 的两个 pivot 命令在每个 HPIn 旁边设置 Median。
您可以使用 facet_grid 或 facet_wrap.
解决方案
library(dplyr)
library(tidyr)
library(ggplot2)
df %>%
# transform in date
mutate(Month = as.Date(paste0("01/", Month), format = "%d/%m/%Y")) %>%
# reshape data
pivot_wider(names_from = Index, values_from = Value) %>%
pivot_longer(starts_with("HPI"), names_to = "Index", values_to = "Value") %>%
# plot by HPI
ggplot(aes(x = Month)) +
geom_line(aes(y = Value, colour = Index)) +
geom_line(aes(y = Median, colour = "Median")) +
geom_point(aes(y = Value, colour = Index)) +
geom_point(aes(y = Median, colour = "Median")) +
scale_x_date(date_labels = "%m %Y", date_breaks = "1 month") +
facet_grid(Index~Operation)
图例是多余的。如果您不想要它：删除第二个 geom_line 中的 color = "Median" 并在第一个 geom_line 中添加 show.legend = FALSE。
或者你可以在最后添加+ theme(legend.position = "none")。
数据
# (I just tripled your data)
df <- tibble::tribble(
~Index, ~Value, ~Operation, ~Month,
"HPI1", 0.9, "Sale", "01/2020",
"HPI1", 1.1, "Rent", "02/2020",
"HPI2", 0.89, "Sale", "01/2020",
"HPI2", 1.12, "Rent", "02/2020",
"HPI3", 0.85, "Sale", "01/2020",
"HPI3", 1.22, "Rent", "02/2020",
"Median", 0.91, "Sale", "01/2020",
"Median", 1.02, "Rent", "02/2020",
"HPI1", 0.9, "Sale", "02/2020",
"HPI1", 1.1, "Rent", "03/2020",
"HPI2", 0.89, "Sale", "02/2020",
"HPI2", 1.12, "Rent", "03/2020",
"HPI3", 0.85, "Sale", "02/2020",
"HPI3", 1.22, "Rent", "03/2020",
"Median", 0.91, "Sale", "02/2020",
"Median", 1.02, "Rent", "03/2020",
"HPI1", 0.9, "Sale", "03/2020",
"HPI1", 1.1, "Rent", "04/2020",
"HPI2", 0.89, "Sale", "03/2020",
"HPI2", 1.12, "Rent", "04/2020",
"HPI3", 0.85, "Sale", "03/2020",
"HPI3", 1.22, "Rent", "04/2020",
"Median", 0.91, "Sale", "03/2020",
"Median", 1.02, "Rent", "04/2020")
另一种使用 foreach 循环的方法
注意：我使用了@Edo answer
中提供的示例数据初始库加载并准备图例供以后使用
library(ggplot2) # For obvious reason - graph
library(dplyr) # For manipulate data a bit
library(foreach) # Foreach Loop
library(cowplot) # For manipulate graphs
# Define the line type Solid for variables and dotted line for Median
line_type <- c("solid", "1111")
names(line_type) <- c("Index (HPI1, HPI2, HPI3)", "Median")
# Generate dummy plot just to extract legend
plot_for_legend <- ggplot(data = RES %>%
mutate(Index = if_else(Index == "Median", "Median",
# If you have more Index than HPI1, HPI2, HPI3, can just simply remove
# the text in parentheses and leave it as Index - make sure you change
# the line_type variable accordingly and adjust all the places needed.
"Index (HPI1, HPI2, HPI3)")),
aes(x = Month, y = Value, group = Index)) +
geom_line(aes(linetype = Index)) +
scale_linetype_manual(values = line_type) +
theme(legend.position = 'bottom', legend.title = element_blank(),
axis.text.x = element_text(angle = 90))
# Extracting legend using get_legend from cowplot package
legend_area <- get_legend(plot_for_legend)
使用 foreach 循环遍历每个操作并绘制它们
operation <- unique(RES[["Operation"]])
operation_plot <- foreach(z = operation) %do% {
# filter data to current operation Rent/ Sale
operation_data <- RES %>% filter(Operation == z)
median_data <- operation_data %>% filter(Index == "Median")
median_line_point <- geom_line(data = median_data,
aes(x = Month, y = Value, group = Index, linetype = "Median"),
alpha=1)
median_point <- geom_point(data = median_data,
aes(x = Month, y = Value))
list_index <- unique(operation_data[["Index"]])
list_index <- list_index[list_index != "Median"]
list_plot <- foreach(i_index = list_index) %do% {
ggplot(data = operation_data %>% filter(Index == i_index),
aes(x = Month, y = Value, group = Index,
# in case you change the line_type make sure to change the name
# here accordingly
linetype ="Index (HPI1, HPI2, HPI3)"),
alpha=1) +
geom_line(linetype = "solid") +
geom_point() +
median_line_point + median_point +
scale_linetype_manual(values = line_type) +
geom_hline(aes(yintercept = 1,colour = "red"), show.legend = F) +
ylab(i_index) +
theme(legend.position = "none")
}
# put all the plot in the list into one
current_plot <- plot_grid(plotlist = list_plot, ncol = 2)
# Generate title for operation
title <- ggdraw() +
draw_label(
sprintf("%s operation", z),
fontface = 'bold',
x = 0,
hjust = 0
) +
theme(
# add margin on the left of the drawing canvas,
# so title is aligned with left edge of first plot
plot.margin = margin(0, 0, 0, 7)
)
# Combine title, main plot & legend into one
current_plot_w_legend <- plot_grid(title, current_plot, legend_area, nrow = 3, rel_heights = c(.5, 10, 1))
}
# Set names of the plot variables for easier access
names(operation_plot) <- operation
这里有一些示例数据是输出