如何通过 dplyr 在 R 上生成频率表并用 ggplot 绘制其值？答案

【问题标题】：How do I generate a frequency table on R via dplyr and plot its values with ggplot?如何通过 dplyr 在 R 上生成频率表并用 ggplot 绘制其值？
【发布时间】：2018-03-06 01:48:03
【问题描述】：

我需要从两个分类变量列中创建一个频率表，其中一个是 5 岁年龄组，另一个是来自 brfss2013 数据集的健康状况（五个州），我从中提取了感兴趣的列：

> hlthgrpq1 <- brfss2013 %>% select(genhlth, X_ageg5yr)

从而生成一个两列框架，对 2 个变量进行 491775 次观察。

    genhlth     X_ageg5yr
1   Fair        Age 60 to 64
2   Good        Age 50 to 54
3   Good        Age 55 to 59
4   Very good   Age 60 to 64
5   Good        Age 65 to 69

我可以用'by'函数生成一个汇总表：

> by(hlthgrpq1$genhlth, hlthgrpq1$X_ageg5yr, summary)
hlthgrpq1$X_ageg5yr: Age 18 to 24
Excellent Very good      Good      Fair      Poor      NA's 
     6896     10266      7795      1873       303        69 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 25 to 29
Excellent Very good      Good      Fair      Poor      NA's 
     5779      8488      6521      1751       325        46 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 30 to 34
Excellent Very good      Good      Fair      Poor      NA's 
     6412      9958      7977      2295       496        75 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 35 to 39
Excellent Very good      Good      Fair      Poor      NA's 
     6366     10169      8236      2637       638        61 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 40 to 44
Excellent Very good      Good      Fair      Poor      NA's 
     6689     11130      9193      3334      1067        95 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 45 to 49
Excellent Very good      Good      Fair      Poor      NA's 
     7051     12278     10611      4343      1815       112 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 50 to 54
Excellent Very good      Good      Fair      Poor      NA's 
     8545     15254     13761      6354      3120       139 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 55 to 59
Excellent Very good      Good      Fair      Poor      NA's 
     8500     16759     15394      7643      3998       197 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 60 to 64
Excellent Very good      Good      Fair      Poor      NA's 
     8283     16825     16266      8101      3955       229 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 65 to 69
Excellent Very good      Good      Fair      Poor      NA's 
     7479     15764     15600      7749      3200       205 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 70 to 74
Excellent Very good      Good      Fair      Poor      NA's 
     5491     11943     13125      6491      2721       196 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 75 to 79
Excellent Very good      Good      Fair      Poor      NA's 
     3320      8501     10128      5545      2426       173 
---------------------------------------------------------------------------------------------------------------- 
hlthgrpq1$X_ageg5yr: Age 80 or older
Excellent Very good      Good      Fair      Poor      NA's 
     3697     10285     14400      8116      3695       322

这就是我卡住的地方。我已经尝试了几个小时试图到达这里：

Results obtained via spreadsheet.

感谢您的帮助。

（这是针对特定任务的，所以我只能使用 dplyr 和 ggplot2，所以，没有 reshape2 或 tidyr。）

【问题讨论】：

看看dplyr动词group_by()和summarise()。

标签： r ggplot2 dplyr

【解决方案1】：

首先：对于以后的发布，最好始终包含示例数据。请参阅here how to provide a minimal reproducible example/attempt，包括示例数据。

基础 R 中的解决方案。

as.data.frame.matrix(t(table(df)));
#             Fair Good Very good
#Age 50 to 54    0    1         0
#Age 55 to 59    0    1         0
#Age 60 to 64    1    0         1
#Age 65 to 69    0    1         0

或者类似tidyverse 的方法？

library(tidyverse);
df %>% count(genhlth, X_ageg5yr) %>% spread(genhlth, n);
## A tibble: 4 x 4
#  X_ageg5yr     Fair  Good `Very good`
#  <fct>        <int> <int>       <int>
#1 Age 50 to 54    NA     1          NA
#2 Age 55 to 59    NA     1          NA
#3 Age 60 to 64     1    NA           1
#4 Age 65 to 69    NA     1          NA

或者如果你坚持只使用dplyr 而不是tidyr，你可以这样做：

df2 <- df %>%
    count(genhlth, X_ageg5yr);
df2 <- as.data.frame.matrix(xtabs(n ~ X_ageg5yr + genhlth, data = df2));
#             Fair Good Very good
#Age 50 to 54    0    1         0
#Age 55 to 59    0    1         0
#Age 60 to 64    1    0         1
#Age 65 to 69    0    1         0

这基本上归结为从宽到长的重新格式化，SO 围绕该主题进行了大量讨论（例如here）。

样本数据

df <- read.table(text =
    "genhlth     X_ageg5yr
Fair        'Age 60 to 64'
Good        'Age 50 to 54'
Good        'Age 55 to 59'
'Very good'   'Age 60 to 64'
Good        'Age 65 to 69'", header = T)

【讨论】：