大量的数据争论与选举数据有关。将行分组为列？

Question

我已经从俄克拉荷马州选举委员会下载了与最近选举有关的数据，并希望将这些数据转化为能够进行更合理分析的数据。不幸的是，数据的格式并没有使这变得容易。

目前，数据如下：

   precinct race_description cand_party cand_tot_votes
1     10001         GOVERNOR        DEM            106
2     10001         GOVERNOR        IND              5
3     10001         GOVERNOR        LIB              7
4     10001         GOVERNOR        REP            118
5     10002         GOVERNOR        DEM             84
6     10002         GOVERNOR        IND              9
7     10002         GOVERNOR        LIB              3
8     10002         GOVERNOR        REP            151
9     10003         GOVERNOR        DEM             36
10    10003         GOVERNOR        IND              2

有什么办法让我把这个缩小到每一个分局都是一行，其他数据就在一列中吗？因此，我会有一排分区号10001的结果，和一个列标题，类似于"Gov Dem"，"Gov Rep"，等等。

老实说，我不知道该怎么处理这个问题。

Answer 1

您可以使用names_sep参数在pivot_wider中为名称组合两列。

library(tidyverse)

df %>% 
  pivot_wider(
    names_from = c(race_description, cand_party),
    names_sep = "_",
    values_from = cand_tot_votes
  )

输出

  precinct GOVERNOR_DEM GOVERNOR_IND GOVERNOR_LIB GOVERNOR_REP
     <int>        <int>        <int>        <int>        <int>
1    10001          106            5            7          118
2    10002           84            9            3          151
3    10003           36            2           NA           NA

如果你想要不同的标题，那么你总是可以在旋转之前清理这些标题：

df %>% 
  mutate(race_description = str_extract(str_to_title(race_description), "^.{3}"),
         cand_party = str_to_title(cand_party)) %>% 
  pivot_wider(
    names_from = c(race_description, cand_party),
    names_sep = "_",
    values_from = cand_tot_votes
  )

#  precinct Gov_Dem Gov_Ind Gov_Lib Gov_Rep
#     <int>   <int>   <int>   <int>   <int>
#1    10001     106       5       7     118
#2    10002      84       9       3     151
#3    10003      36       2      NA      NA

数据

df <- structure(list(precinct = c(10001L, 10001L, 10001L, 10001L, 10002L, 
10002L, 10002L, 10002L, 10003L, 10003L), race_description = c("GOVERNOR", 
"GOVERNOR", "GOVERNOR", "GOVERNOR", "GOVERNOR", "GOVERNOR", "GOVERNOR", 
"GOVERNOR", "GOVERNOR", "GOVERNOR"), cand_party = c("DEM", "IND", 
"LIB", "REP", "DEM", "IND", "LIB", "REP", "DEM", "IND"), cand_tot_votes = c(106L, 
5L, 7L, 118L, 84L, 9L, 3L, 151L, 36L, 2L)), class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6", "7", "8", "9", "10"))