如何使用tidycensus API合并三州种族数据为数据框并通过spread函数生成种族列

先修正两个关键问题 #

1. merge()函数仅支持两两合并，无法一次传入三个数据框，这是你报错的直接原因。
2. 你的代码最后用get_acs()重新赋值了ny变量，把之前获取的十年普查种族数据完全覆盖了，后续合并会丢失原始数据，必须给不同来源的数据单独命名。

步骤1：重新获取并命名三州种族数据 #

给每个州的种族数据单独命名，避免覆盖：

racevars <- c(White = "P2_005N", 
              Black = "P2_006N", 
              Asian = "P2_008N", 
              Hispanic = "P2_002N")

# NY种族数据
ny_race <- get_decennial(
  geography = "tract",
  variables = racevars,
  state = "NY",
  geometry = TRUE,
  summary_var = "P2_001N",
  year = 2020
) 

# NJ种族数据
nj_race <- get_decennial(
  geography = "tract",
  variables = racevars,
  state = "NJ",
  geometry = TRUE,
  summary_var = "P2_001N",
  year = 2020
) 

# CT种族数据
ct_race <- get_decennial(
  geography = "tract",
  variables = racevars,
  state = "CT",
  geometry = TRUE,
  summary_var = "P2_001N",
  year = 2020
) 

步骤2：合并三州数据 #

因为三个数据框结构完全一致，用dplyr::bind_rows()做行合并，同时添加state列标记数据所属州：

library(dplyr)
library(tidyr)

total_race <- bind_rows(
  ny_race %>% mutate(state = "NY"),
  nj_race %>% mutate(state = "NJ"),
  ct_race %>% mutate(state = "CT")
)

步骤3：用spread（或推荐的pivot_wider）生成种族单独列 #

用spread（旧版tidyr函数）：

# 先计算种族人口占比（按需选择）
total_race <- total_race %>%
  mutate(percent = 100 * (value / summary_value))

# 转成宽表，每个种族对应一列
total_race_wide <- total_race %>%
  select(GEOID, NAME, state, variable, percent, geometry) %>% # 保留需要的列
  spread(key = variable, value = percent)

用pivot_wider（tidyr 1.0.0+推荐函数）：

total_race_wide <- total_race %>%
  mutate(percent = 100 * (value / summary_value)) %>%
  select(GEOID, NAME, state, variable, percent, geometry) %>%
  pivot_wider(names_from = variable, values_from = percent)

内容的提问来源于stack exchange，提问作者WutsKraken