绘制具有不同基本总体的 R 中的百分比

Question

我想用 ggplot2 包直观地比较 R 中美国两个州的交通站点的两个数据集。我将它们组合到一个数据框中，显示每年和州的交通站点总数。由于这些数字非常不同，我想比较每个州的人口百分比。首先，这是一个示例 df 以及我到目前为止所取得的成就。我在代码中使用了tidyverse 和lubridate。

df <- data.frame(ID=c("CA-2013-0000001","CA-2014-0000001", "TX-2013-0000001", "TX-2014-0000001"),
                    State=c("CA", "CA", "TX", "TX"),
                    Stop_Date=ymd("2013-01-01","2014-01-01", "2013-01-01", "2014-01-01"))

df %>%
  group_by(year = year(Stop_Date), state = State) %>%
  count() %>%
  ggplot(aes(year, n, col = state))+
  geom_point(stat = "identity")+
  geom_line(stat = "identity")

使用该代码，我得到了一个有两条线的图，每条线都反映了我看到的两种状态。

我想创建完全相同的图，但我想显示与州人口相关的百分比，而不是总数，即population_ca <- 38620000 和population_tx <- 26980000。

我尝试了这两种方法，但每次运行代码时都会返回不同的错误：

df %>%
  group_by(year = year(Stop_Date), state = State) %>%
  summarise(PercentStopsToPopulation = if_else(state == "CA",
                                                 ((n()/population_ca)*100),
                                                 ((n()/population_tx)*100))) %>%
  ggplot(aes(year, PercentStopsToPopulation, col = state))+
  geom_point(stat = "identity")+
  geom_line(stat = "identity")

df %>%
  group_by(year = year(Stop_Date), state = State) %>%
  summarise(PercentCA = ifelse(state == "CA",((n()/population_ca)*100)),
            PercentTX = ifelse(state == "TX", ((n()/population_tx)*100))) %>%
  ggplot(aes(year, PercentCA))+
  geom_point(stat = "identity")+
  geom_line(stat = "identity")+
  geom_point(aes(year, PercentTX), stat = "identity")+
  geom_line(aes(year, PercentTX), stat = "identity")

我真的希望有人可以帮助我并告诉我我的错误在哪里。 提前谢谢！

Answer 1

以下是我解决此问题的方法。我使用 tidyverse，所以你会注意到一些变化。

library("tidyverse")

#keep organized and avoid for loops by organizing population data in a tibble
pop <- tibble(state = c("CA", "TX"),
              population = c(38620000, 26980000))

#I made this a tibble instead of a dataframe, just to stay consistent in the tidyverse approach.
df <- tibble(ID=c("CA-2013-0000001","CA-2014-0000001", "TX-2013-0000001", "TX-2014-0000001"),
                 state=c("CA", "CA", "TX", "TX"),
                 stop_date=ymd("2013-01-01","2014-01-01", "2013-01-01", "2014-01-01")) %>%
  #I prefer to err on the side of making more fields, to make it easier to see what we're doing down the road.
  mutate(year = year(stop_date))

# summarise data
df_count <- df %>%
  group_by(year, state) %>%
  count() %>%
  #Join with population table. I prefer this over a for loop - easier to scale up, in case you decide to add more states.
  full_join(pop) %>%
  # Calculate the percent of population
  mutate(percent = 100*n/population)


#Now, we graph!
df_count %>%
  ggplot(aes(year, percent, col = state))+
  geom_point()+
  geom_line()

如果您有任何问题，请告诉我。 :)