Data Challenge Lab Home

Exploratory data analysis (1D) [explore]

(Builds on: Visualization basics, Manipulation basics)
(Leads to: Exploratory data analysis (2D), Function basics, Essentials of relational data, String basics, Tidy data)

Exploratory data analysis is partly a set of techniques, but is mostly a mindset: you want to remain open to what the data is telling you.

library(tidyverse)
library(nycflights13)

Whenever you start working with a new variable, it’s a really good idea to first take a look at the variable by itself, before you start combining it with other variables. As well as the visual techniques you’ll learn in the readings, another quick and dirty function is count().

df %>% count(grp) is shorthand for df %>% group_by(grp) %>% summarise(n = n()).

flights %>% 
  count(carrier)
#> # A tibble: 16 x 2
#>    carrier     n
#>    <chr>   <int>
#>  1 9E      18460
#>  2 AA      32729
#>  3 AS        714
#>  4 B6      54635
#>  5 DL      48110
#>  6 EV      54173
#>  7 F9        685
#>  8 FL       3260
#>  9 HA        342
#> 10 MQ      26397
#> 11 OO         32
#> 12 UA      58665
#> 13 US      20536
#> 14 VX       5162
#> 15 WN      12275
#> 16 YV        601

It has two convenient arguments:

You can also count() the value of expression. This is a useful technique to get a quick count of how many missing values there are:

flights %>% 
  count(is.na(dep_delay))
#> # A tibble: 2 x 2
#>   `is.na(dep_delay)`      n
#>   <lgl>               <int>
#> 1 FALSE              328521
#> 2 TRUE                 8255

flights %>% 
  count(
    dep_missing = is.na(dep_time), 
    arr_missing = is.na(arr_time)
  )
#> # A tibble: 3 x 3
#>   dep_missing arr_missing      n
#>   <lgl>       <lgl>        <int>
#> 1 FALSE       FALSE       328063
#> 2 FALSE       TRUE           458
#> 3 TRUE        TRUE          8255

You can combine count() with the cut_* functions from ggplot2 to compute histograms “by hand”:

# five bins of equal widths
flights %>% 
  count(cut_interval(arr_delay, 5))
#> # A tibble: 6 x 2
#>   `cut_interval(arr_delay, 5)`      n
#>   <fct>                         <int>
#> 1 [-86,186]                    323807
#> 2 (186,457]                      3465
#> 3 (457,729]                        45
#> 4 (729,1e+03]                      25
#> 5 (1e+03,1.27e+03]                  4
#> 6 <NA>                           9430

# five bins with approximately equal numbers of points
flights %>% 
  count(cut_number(arr_delay, 5))
#> # A tibble: 6 x 2
#>   `cut_number(arr_delay, 5)`     n
#>   <fct>                      <int>
#> 1 [-86,-19]                  70875
#> 2 (-19,-10]                  61570
#> 3 (-10,1]                    66972
#> 4 (1,21]                     62970
#> 5 (21,1.27e+03]              64959
#> 6 <NA>                        9430

# hourly bins
flights %>% 
  count(cut_width(arr_delay, 60, boundary = 0))
#> # A tibble: 22 x 2
#>    `cut_width(arr_delay, 60, boundary = 0)`      n
#>    <fct>                                     <int>
#>  1 [-120,-60]                                  240
#>  2 (-60,0]                                  194102
#>  3 (0,60]                                   105215
#>  4 (60,120]                                  17755
#>  5 (120,180]                                  6191
#>  6 (180,240]                                  2291
#>  7 (240,300]                                   941
#>  8 (300,360]                                   365
#>  9 (360,420]                                   144
#> 10 (420,480]                                    37
#> # … with 12 more rows

Readings