suggestion for `shadow_shift` and `label_missing`

[njtierney]

suggestion for `shadow_shift` and `label_missing`

[njtierney]

Something similar to this can be achieved with add_shadow_shift, except for the miss_label.

library(narnia)
library(tidyverse)

airquality %>%
  add_shadow_shift(vars = c("Ozone", "Solar.R")) %>%
  mutate(miss_label = label_missing_2d(Ozone,Solar.R)) %>%
  head()
#> # A tibble: 6 x 9
#>   Ozone Solar.R  Wind  Temp Month   Day Ozone_shift Solar.R_shift
#>   <int>   <int> <dbl> <int> <int> <int>       <dbl>         <dbl>
#> 1    41     190   7.4    67     5     1    41.00000     190.00000
#> 2    36     118   8.0    72     5     2    36.00000     118.00000
#> 3    12     149  12.6    74     5     3    12.00000     149.00000
#> 4    18     313  11.5    62     5     4    18.00000     313.00000
#> 5    NA      NA  14.3    56     5     5   -14.79436     -25.09977
#> 6    28      NA  14.9    66     5     6    28.00000     -19.64529
#> # ... with 1 more variables: miss_label <chr>

It would be interesting to add shadow cols onto each variable added, and then a catch all label.
Perhaps there might be a more compact way of showing the below code:

airquality %>%
  select(Ozone, Solar.R) %>%
  add_shadow_shift(vars = c("Ozone", "Solar.R")) %>%
  cast_shadow(vars = c("Ozone", "Solar.R")) %>%
  select(-Ozone,-Solar.R) %>%
  mutate(any_missing = label_missings(airquality))
#> # A tibble: 153 x 5
#>    Ozone_shift Solar.R_shift Ozone_NA Solar.R_NA any_missing
#>          <dbl>         <dbl>   <fctr>     <fctr>       <chr>
#>  1    41.00000     190.00000      !NA        !NA Not Missing
#>  2    36.00000     118.00000      !NA        !NA Not Missing
#>  3    12.00000     149.00000      !NA        !NA Not Missing
#>  4    18.00000     313.00000      !NA        !NA Not Missing
#>  5   -14.94837     -28.02921       NA         NA     Missing
#>  6    28.00000     -32.93632      !NA         NA     Missing
#>  7    23.00000     299.00000      !NA        !NA Not Missing
#>  8    19.00000      99.00000      !NA        !NA Not Missing
#>  9     8.00000      19.00000      !NA        !NA Not Missing
#> 10   -16.26483     194.00000       NA        !NA     Missing
#> # ... with 143 more rows

[njtierney]

I think that the best I can do now is this:

library(narnia)
library(tidyverse)
#> Loading tidyverse: ggplot2
#> Loading tidyverse: tibble
#> Loading tidyverse: tidyr
#> Loading tidyverse: readr
#> Loading tidyverse: purrr
#> Loading tidyverse: dplyr
#> Conflicts with tidy packages ----------------------------------------------
#> filter(): dplyr, stats
#> lag():    dplyr, stats
aq_shift <- airquality %>%
  cast_shadow_shift(vars = c("Ozone", "Solar.R")) %>%
  add_label_missings()

aq_shift
#> # A tibble: 153 x 7
#>    Ozone Solar.R Ozone_NA Solar.R_NA Ozone_shift Solar.R_shift any_missing
#>    <int>   <int>   <fctr>     <fctr>       <dbl>         <dbl>       <chr>
#>  1    41     190      !NA        !NA    41.00000     190.00000 Not Missing
#>  2    36     118      !NA        !NA    36.00000     118.00000 Not Missing
#>  3    12     149      !NA        !NA    12.00000     149.00000 Not Missing
#>  4    18     313      !NA        !NA    18.00000     313.00000 Not Missing
#>  5    NA      NA       NA         NA   -17.38178     -29.47859     Missing
#>  6    28      NA      !NA         NA    28.00000     -33.72173     Missing
#>  7    23     299      !NA        !NA    23.00000     299.00000 Not Missing
#>  8    19      99      !NA        !NA    19.00000      99.00000 Not Missing
#>  9     8      19      !NA        !NA     8.00000      19.00000 Not Missing
#> 10    NA     194       NA        !NA   -11.54721     194.00000     Missing
#> # ... with 143 more rows

  ggplot(aq_shift,
         aes(x = Ozone_shift,
             y = Solar.R_shift,
             colour = any_missing)) + 
  geom_point()

This allows the user to get the same data structure out that powers geom_missing_point()

library(narnia)
  library(tidyverse)
#> Loading tidyverse: ggplot2
#> Loading tidyverse: tibble
#> Loading tidyverse: tidyr
#> Loading tidyverse: readr
#> Loading tidyverse: purrr
#> Loading tidyverse: dplyr
#> Conflicts with tidy packages ----------------------------------------------
#> filter(): dplyr, stats
#> lag():    dplyr, stats
  ggplot(airquality,
         aes(x = Ozone,
             y = Solar.R)) + 
    geom_missing_point()

[njtierney]

I have shortened this code to be a little more concise, with the (slightly verbose) function cast_shadow_shift_label

library(tidyverse)
#> Loading tidyverse: ggplot2
#> Loading tidyverse: tibble
#> Loading tidyverse: tidyr
#> Loading tidyverse: readr
#> Loading tidyverse: purrr
#> Loading tidyverse: dplyr
#> Conflicts with tidy packages ----------------------------------------------
#> filter(): dplyr, stats
#> lag():    dplyr, stats
library(narnia)

# using cast is like transmute - it just casts shadow vars for the 
# variables of interest, this facilitates plotting and other summaries
aq_shift <- airquality %>% cast_shadow_shift_label(c("Ozone", "Solar.R"))

aq_shift
#> # A tibble: 153 x 7
#>    Ozone Solar.R Ozone_NA Solar.R_NA Ozone_shift Solar.R_shift any_missing
#>    <int>   <int>   <fctr>     <fctr>       <dbl>         <dbl>       <chr>
#>  1    41     190      !NA        !NA    41.00000     190.00000 Not Missing
#>  2    36     118      !NA        !NA    36.00000     118.00000 Not Missing
#>  3    12     149      !NA        !NA    12.00000     149.00000 Not Missing
#>  4    18     313      !NA        !NA    18.00000     313.00000 Not Missing
#>  5    NA      NA       NA         NA   -17.13731     -17.57498     Missing
#>  6    28      NA      !NA         NA    28.00000     -32.76235     Missing
#>  7    23     299      !NA        !NA    23.00000     299.00000 Not Missing
#>  8    19      99      !NA        !NA    19.00000      99.00000 Not Missing
#>  9     8      19      !NA        !NA     8.00000      19.00000 Not Missing
#> 10    NA     194       NA        !NA   -14.16446     194.00000     Missing
#> # ... with 143 more rows

ggplot(aq_shift,
       aes(x = Ozone_shift,
           y = Solar.R_shift,
           colour = any_missing)) + 
  geom_point()

[njtierney]

Done!