在 R 中使用 dplyr 重塑 table

Question

欢迎就 dplyr 在 R 中的正确应用提出一些建议。我们有以下数据：

   City            Amount    Category
1  Los Angeles     100       Film
2  Los Angeles     200       Film
3  Los Angeles     400       Music 
4  Seattle         300       Coffee
5  Boston          600       Books
...

最终结果应如下所示：

                        Film   Coffee   Books   ...
City  
Los Angeles, CA         Sum    Sum      Sum     Sum 
Seattle, WA             Sum    Sum      Sum     Sum 
Boston, MA              Sum    Sum      Sum     Sum

我希望数据透视表 table 汇总每个城市中每个类别的 "Amount" 的总值，以便城市在一列中位于左侧，所有类别在顶部作为一行.

尝试过：

data %>%                                            
  group_by(Location, Category) %>%
  summarise(Amount = sum(Amount))

哪个看起来更像

   City            Amount    Category
1  Los Angeles     300       Film
3  Los Angeles     400       Music 
4  Seattle         300       Coffee
5  Boston          600       Books

计算是正确的，但如前所述，我们需要将城市和类别作为矩阵，其中每个单元格内的每个金额之和。

感谢您的帮助！

Answer 1

您正在寻找的是 tidyr::spread 将您的 data.frame 从长格式重塑为宽格式：

library(tidyverse)

# recreate the data
data <- tribble(
  ~City,             ~Amount,   ~Category,
  "Los Angeles",     100,       "Film",
  "Los Angeles",     200,       "Film",
  "Los Angeles",     400,       "Music", 
  "Seattle",         300,       "Coffee",
  "Boston",          600,       "Books"
)

# using your code to get the data in the long-format
data_long <- data %>% 
  group_by(City, Category) %>%
  summarise(Amount = sum(Amount))

data_long
#> # A tibble: 4 x 3
#> # Groups:   City [?]
#>          City Category Amount
#>         <chr>    <chr>  <dbl>
#> 1      Boston    Books    600
#> 2 Los Angeles     Film    300
#> 3 Los Angeles    Music    400
#> 4     Seattle   Coffee    300

# spread to wide using the tidyr-package (in tidyverse)
data_wide <- spread(data_long, key = "Category", value = "Amount", fill = 0)

data_wide
#> # A tibble: 3 x 5
#> # Groups:   City [3]
#>          City Books Coffee  Film Music
#> *       <chr> <dbl>  <dbl> <dbl> <dbl>
#> 1      Boston   600      0     0     0
#> 2 Los Angeles     0      0   300   400
#> 3     Seattle     0    300     0     0

走向矩阵

mat <- as.matrix(data_wide %>% ungroup %>% select(-City))
rownames(mat) <- data_wide$City

mat
#>             Books Coffee Film Music
#> Boston        600      0    0     0
#> Los Angeles     0      0  300   400
#> Seattle         0    300    0     0

str(mat)
#>  num [1:3, 1:4] 600 0 0 0 0 300 0 300 0 0 ...
#>  - attr(*, "dimnames")=List of 2
#>   ..$ : chr [1:3] "Boston" "Los Angeles" "Seattle"
#>   ..$ : chr [1:4] "Books" "Coffee" "Film" "Music"

在 R 中使用 dplyr 重塑 table

Reshaping a table with dplyr in R

r

transform

reshape

dplyr

走向矩阵