Distance Function

[spsanderson]

Distance Function(s)

Vector Function

The vector distance() function should compute the point to point distance of any one of 1D, 2D or 3D Random Walks. Each function has attributes that will indicate it's dimension. For example:

library(RandomWalker)
attributes(rw30())
$names
[1] "walk_number" "x"           "y"          

$row.names
   [1]    1    2    3    4    5   ...

$class
[1] "tbl_df"     "tbl"        "data.frame"

$num_walks
[1] 30

$num_steps
[1] 100

$mu
[1] 0

$sd
[1] 1

$fns
[1] "rw30"

$dimension
[1] 1

This would allow any function to make use of any of the underlying information. Here we would choose $dimension to see that it is a 1, 2 or3D random walk.

One and Two Dimensional Random Walks

I think these should work in the following manner:

1. User creates data with something like rw30() and pipes it out to distance()
2. distance will create a vector of point to point distances traveled.
3. This distance() function would be the basis for a distance_augment() function which would augment distance to the passed data.frame/tibble
4. A parameter of .y can default to y but can take an argument of a column name to pass. We would only take the distance of the randomly generated variable y as a default, otherwise for 1D and 2D walks .y can be specified like cum_sum or cum_prod

This introduces a problem of how should the distance function be written, should there be multiple .y parameters or should an ... argument be provided to a select function like dpyr::select(...), or, one could provide a list of columns like .y = c(y1, y2, y3)

Here is some sample code:

library(RandomWalker)

df <- rw30()
head(df)
# A tibble: 6 × 3
  walk_number     x     y
  <fct>       <int> <dbl>
1 1               1  0   
2 1               2 -2.35
3 1               3 -2.31
4 1               4 -3.43
5 1               5 -5.50
6 1               6 -5.20
 
df |>
  dplyr::mutate(
    distance = sqrt((x - dplyr::lag(x,1))^2 + (y - dplyr::lag(y,1))^2)
  )
# A tibble: 3,000 × 4
   walk_number     x     y distance
   <fct>       <int> <dbl>    <dbl>
 1 1               1  0       NA   
 2 1               2 -2.35     2.56
 3 1               3 -2.31     1.00
 4 1               4 -3.43     1.50
 5 1               5 -5.50     2.30
 6 1               6 -5.20     1.05
 7 1               7 -6.18     1.40
 8 1               8 -5.45     1.24
 9 1               9 -4.13     1.65
10 1              10 -3.72     1.08
# ℹ 2,990 more rows
# ℹ Use `print(n = ...)` to see more rows

[spsanderson]

A 1D distance function is operational.