r_eda

占比选图

参考 Strayer (2019)

对于占比的表示，一共有三种图表可以展示

1. 圆饼状图，pie，用角度度量
2. 华夫饼状图，waffle，用面积度量
3. 累计直方图，用部分面积度量，取极坐标就是圆饼状图（参考相关 ggplot2 的代码）

当数据的类别超过3个时，圆饼状图很难进行比较。

Intuitively, you can think about a pie chart as a stacked bar chart that has been ‘wrapped’ around some central axis. DataCamp

pie

圆饼图和累计直方图类似。 在 ggplot2 中只需要两个函数

1. coord_polar(theta = 'y')
2. x = 1

suppressMessages(library(tidyverse))

who_disease <- read_csv('datasets/who_disease.csv')

## Parsed with column specification:
## cols(
##   region = col_character(),
##   countryCode = col_character(),
##   country = col_character(),
##   disease = col_character(),
##   year = col_double(),
##   cases = col_double()
## )

# Wrangle data into form we want. 
disease_counts <- who_disease %>%
    mutate(disease = ifelse(disease %in% c('measles', 'mumps'), disease, 'other')) %>%
    group_by(disease) %>%
    summarise(total_cases = sum(cases))

pie_raw <- 
    ggplot(disease_counts, aes(x = 1, y = total_cases, fill = disease)) +
    # Use a column geometry.
    geom_col() + 
    # Change coordinate system to polar and set theta to 'y'.
    coord_polar(theta = 'y') 
pie_raw

可见 pie 不美观，设计中间变量 pie_raw 进行美化。

pie_raw +
    theme_void() +
    # 做一定清洗
    labs(title = 'Proportion of diseases')

waffle

由上可知，当 class 变多，pie 就不适用了。

disease_counts <- who_disease %>%
    group_by(disease) %>%
    summarise(total_cases = sum(cases)) %>% 
    mutate(percent = round(total_cases/sum(total_cases)*100))

# Create an array of rounded percentages for diseases.
case_counts <- disease_counts$percent
# Name the percentage array with disease_counts$disease
names(case_counts) <- disease_counts$disease

# Pass case_counts vector to the waffle function to plot
library(waffle)
waffle(case_counts)

注意这里 waffle 的 input 是

parts named vector of values to use for the chart

是一个已经命名的向量。

stacked bar

当比例除了需要组内比较，还需要组间比较，就需要累计直方图了。

disease_counts <- who_disease %>%
    mutate(
        disease = ifelse(disease %in% c('measles', 'mumps'), disease, 'other') %>% 
        factor(levels = c('measles', 'other', 'mumps')) # change factor levels to desired ordering
    ) %>%
    group_by(disease, year) %>%
    summarise(total_cases = sum(cases)) 

# plot
ggplot(disease_counts, aes(x = year, y = total_cases, fill = disease)) +
    geom_col(position = 'fill')

当调整了分类变量的 levels 后，可以主要观察 measles 和 munmps 的占比组间变化了。

Just remember to keep the number of classes low to aid readibility! DataCamp

因此这也暴露了一个问题，bar 和 pie 都不能容纳太多的 levels。

Strayer, Nick. 2019. “Visualization Best Practices in R.” DataCamp. 2019. <https://www.datacamp.com/courses/visualization-best-practices-in-r>.