Given a data frame such as might be returned by tidy_draws() or spread_draws(), gather variables and their values from that data frame into a ".variable" and ".value" column.

gather_variables(data, exclude = c(".chain", ".iteration", ".draw", ".row"))

Arguments

data

A data frame with variable names spread across columns, such as one returned by tidy_draws() or spread_draws().

exclude

A character vector of names of columns to be excluded from the gather. Default ignores several meta-data column names used in tidybayes.

Value

A data frame.

Details

This function gathers every column except grouping columns and those matching the expression exclude into key/value columns ".variable" and ".value".

Imagine a data frame data as returned by spread_draws(fit, a[i], b[i,v]), like this:

  • column ".chain": the chain number

  • column ".iteration": the iteration number

  • column ".draw": the draw number

  • column "i": value in 1:5

  • column "v": value in 1:10

  • column "a": value of "a[i]" for draw number ".draw"

  • column "b": value of "b[i,v]" for draw number ".draw"

gather_variables(data) on that data frame would return a grouped data frame (grouped by i and v), with:

  • column ".chain": the chain number

  • column ".iteration": the iteration number

  • column ".draw": the draw number

  • column "i": value in 1:5

  • column "v": value in 1:10

  • column ".variable": value in c("a", "b").

  • column ".value": value of "a[i]" (when ".variable" is "a"; repeated for every value of "v") or "b[i,v]" (when ".variable" is "b") for draw number ".draw"

In this example, this call:

Is roughly equivalent to:

data %>%
  gather(.variable, .value, -c(.chain, .iteration, .draw, i, v)) %>%
  group_by(.variable, .add = TRUE)

Author

Matthew Kay

Examples

# \donttest{

library(dplyr)

data(RankCorr, package = "ggdist")

RankCorr %>%
  spread_draws(b[i,v], tau[i]) %>%
  gather_variables() %>%
  median_qi()
#> # A tibble: 24 x 9
#>        i     v .variable .value  .lower .upper .width .point .interval
#>    <int> <int> <chr>      <dbl>   <dbl>  <dbl>  <dbl> <chr>  <chr>    
#>  1     1     1 b         -1.08  -1.27   -0.897   0.95 median qi       
#>  2     1     1 tau        6.03   5.03    7.11    0.95 median qi       
#>  3     1     2 b         -1.83  -2.08   -1.54    0.95 median qi       
#>  4     1     2 tau        6.03   5.03    7.11    0.95 median qi       
#>  5     1     3 b          0.175  0.0447  0.322   0.95 median qi       
#>  6     1     3 tau        6.03   5.03    7.11    0.95 median qi       
#>  7     1     4 b         -0.101 -0.347   0.134   0.95 median qi       
#>  8     1     4 tau        6.03   5.03    7.11    0.95 median qi       
#>  9     2     1 b         -0.739 -1.01   -0.430   0.95 median qi       
#> 10     2     1 tau        3.30   2.41    4.46    0.95 median qi       
#> # ... with 14 more rows

# the first three lines below are roughly equivalent to ggmcmc::ggs(RankCorr)
RankCorr %>%
  tidy_draws() %>%
  gather_variables() %>%
  median_qi()
#> # A tibble: 19 x 7
#>    .variable  .value  .lower  .upper .width .point .interval
#>    <chr>       <dbl>   <dbl>   <dbl>  <dbl> <chr>  <chr>    
#>  1 b[1,1]    -1.08   -1.27   -0.897    0.95 median qi       
#>  2 b[1,2]    -1.83   -2.08   -1.54     0.95 median qi       
#>  3 b[1,3]     0.175   0.0447  0.322    0.95 median qi       
#>  4 b[1,4]    -0.101  -0.347   0.134    0.95 median qi       
#>  5 b[2,1]    -0.739  -1.01   -0.430    0.95 median qi       
#>  6 b[2,2]     0.0909 -0.384   0.602    0.95 median qi       
#>  7 b[2,3]    -0.175  -0.410   0.0563   0.95 median qi       
#>  8 b[2,4]     0.270  -0.151   0.710    0.95 median qi       
#>  9 b[3,1]    -0.340  -0.609  -0.0615   0.95 median qi       
#> 10 b[3,2]    -0.732  -1.18   -0.286    0.95 median qi       
#> 11 b[3,3]     0.0636 -0.163   0.312    0.95 median qi       
#> 12 b[3,4]     0.0918 -0.327   0.463    0.95 median qi       
#> 13 tau[1]     6.03    5.03    7.11     0.95 median qi       
#> 14 tau[2]     3.30    2.41    4.46     0.95 median qi       
#> 15 tau[3]     3.65    2.73    4.72     0.95 median qi       
#> 16 typical_r  0.548   0.309   0.778    0.95 median qi       
#> 17 u_tau[1]   5.62    3.86    8.03     0.95 median qi       
#> 18 u_tau[2]   5.48    3.46    9.07     0.95 median qi       
#> 19 u_tau[3]   4.96    3.10    7.86     0.95 median qi       

# }