Tidily summarizing multiple metric attributes of artefacts

tidy-multiple-summaries.R

# Summarising multiple metric attributes of artefacts

p1 <- read_excel("data/riumailuoi.xlsx", sheet = 'phase1')

summary_stats_table <- function(x) {
  x %>% 
    select(Mass, 
           Length,  
           Width,     
           Thickness,     
           Scars) %>% 
    filter_all(all_vars(!is.na(.))) %>% 
    summarise_all(c("min", 
                    "mean", 
                    "max", 
                    "median", 
                    "IQR")) %>% 
    gather(key = "key", 
           value = "value") %>%
    separate(key, 
             into = c("measurement", 
                      "statistic"), 
             sep = "[_]") %>% 
    spread(statistic, 
           value)
  
}

summary_stats_table(p1)

# A tibble: 5 x 6
  measurement    IQR     max   mean median   min
  <chr>        <dbl>   <dbl>  <dbl>  <dbl> <dbl>
1 Length        5.05   18.0   12.2   11.3   7.74
2 Mass        450    1500    525    385    80   
3 Scars         7      37     15     12     7   
4 Thickness     1.03    4.04   2.76   2.52  1.21
5 Width         2.34    9.9    7.04   6.88  4.33

# or we can summarise more compactly:

summary_stats_table(p1) %>% 
  mutate(`median (IQR)` = str_glue('{median} ({round(IQR,2)})')) %>% 
  select(measurement, `median (IQR)`)
  
# A tibble: 5 x 2
  measurement `median (IQR)`
  <chr>       <S3: glue>    
1 Length      11.315 (5.05) 
2 Mass        385 (450)     
3 Scars       12 (7)        
4 Thickness   2.52 (1.03)   
5 Width       6.875 (2.34) 

# replace all NA by zero: mutate_all(coalesce, 0)