Add feat_diff_summary() functions to help summarise diff(). Useful …

…for exploring the time gaps in the `index`. (#100)

NAMESPACE

@@ -10,6 +10,14 @@ export("%>%")
 export(add_key_slope)
 export(add_n_obs)
 export(as_tsibble)
+export(b_diff_max)
+export(b_diff_mean)
+export(b_diff_median)
+export(b_diff_min)
+export(b_diff_q25)
+export(b_diff_q75)
+export(b_diff_sd)
+export(b_diff_var)
 export(b_iqr)
 export(b_mad)
 export(b_max)
@@ -26,6 +34,7 @@ export(decreasing)
 export(facet_sample)
 export(facet_strata)
 export(feat_brolgar)
+export(feat_diff_summary)
 export(feat_five_num)
 export(feat_monotonic)
 export(feat_ranges)

NEWS.md

@@ -1,5 +1,6 @@
 # brolgar 0.0.6.9000
 
+* Add `feat_diff_summary()` functions to help summarise diff(). Useful for exploring the time gaps in the `index`. (#100)
 * sample functions now work with multiple keys (#85, #89) (Thanks to @earowang and @deanmarchiori for their help with this.)
 * `facet_sample()` now has a default of 3 per plot
 * resolve features(data ,.key, n_obs) error (#71) 

R/b_summaries.R

@@ -16,7 +16,12 @@
 #'     * b_var: The variance
 #'     * b_mad: The mean absolute deviation
 #'     * b_iqr: The Inter-quartile range
-#' 
+#'     * b_diff_var: The variance diff()
+#'     * b_diff_sd: The standard deviation of diff()
+#'     * b_diff_mean: The mean of diff()
+#'     * b_diff_median: The median of diff()
+#'     * b_diff_q25: The q25 of diff()
+#'     * b_diff_q75: The q75 of diff()
 #' 
 #' @param x a vector
 #' @param na.rm whether to remove NA values. Default is TRUE
@@ -105,6 +110,54 @@ b_iqr <- function(x, na.rm = TRUE, ... ){
   stats::IQR(x, na.rm = na.rm, type = 8, ...)
 }
 
+#' @name b_summaries
+#' @export
+b_diff_var <- function(x, na.rm = TRUE, ...){
+  stats::var(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_sd <- function(x, na.rm = TRUE, ...){
+  stats::sd(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_mean <- function(x, na.rm = TRUE, ...){
+  mean(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_median <- function(x, na.rm = TRUE, ...){
+  stats::median(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_q25 <- function(x, na.rm = TRUE, ...){
+  b_q25(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_q75 <- function(x, na.rm = TRUE, ...){
+  b_q75(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_max <- function(x, na.rm = TRUE, ...){
+  b_max(diff(x, na.rm = na.rm, ...))
+}
+
+#' @name b_summaries
+#' @export
+b_diff_min <- function(x, na.rm = TRUE, ...){
+  b_min(diff(x, na.rm = na.rm, ...))
+}
+
 #  * `l_n_obs()` Number of observations
 #  * `l_slope()` Slope and intercept (given some linear model formula)
 

R/data-world-height.R

@@ -32,6 +32,8 @@
 #'            group = country)) + 
 #'     geom_line()
 #'     
+#' \dontrun{
+#'     
 #' # Explore all samples with `facet_strata()`
 #' ggplot(heights,
 #'        aes(x = year,
@@ -51,5 +53,6 @@
 #' # explore the five number summary of height_cm with `features`
 #' heights %>% 
 #'   features(height_cm, feat_five_num)
+#' }
 #'   
 "heights"

R/facet-sample.R

@@ -21,13 +21,15 @@
 #'   geom_line() +
 #'   facet_sample()
 #' 
+#' \dontrun{
 #' ggplot(heights,
 #'        aes(x = year,
 #'            y = height_cm,
 #'            group = country)) +
 #'   geom_line() +
 #'   facet_sample(n_per_facet = 1,
 #'                n_facets = 12)
+#' }
 facet_sample <- function(n_per_facet = 3,
                          n_facets = 12,
                          nrow = NULL, 

R/facet-strata.R

@@ -19,6 +19,8 @@
 #'            group = country)) +
 #'   geom_line() +
 #'   facet_strata()
+#'   
+#' \dontrun{
 #' 
 #' ggplot(heights,
 #'        aes(x = year,
@@ -43,6 +45,7 @@
 #'   geom_line(aes(group = id)) +
 #'   geom_smooth(method = "lm") + 
 #'   facet_strata(along = .slope_xp)
+#' }
 
 facet_strata <- function(n_strata = 12,
                          along = NULL,

R/features.R

@@ -98,3 +98,17 @@ feat_brolgar <- function(x, ...){
   )
 }
 
+#' @rdname brolgar-features
+#' @export
+feat_diff_summary <- function(x, ...){
+  c(
+  diff_var = b_diff_var(x, ...),
+  diff_sd = b_diff_sd(x, ...),
+  diff_min = b_diff_min(x, ...),
+  diff_q25 = b_diff_q25(x, ...),
+  diff_mean = b_diff_mean(x, ...),
+  diff_median = b_diff_median(x, ...),
+  diff_q75 = b_diff_q75(x, ...),
+  diff_max = b_diff_max(x, ...)
+  )
+}

R/monotonics.R

@@ -33,6 +33,7 @@
 #' unvarying(vec_ran)
 #' unvarying(vec_flat)
 #' 
+#' \dontrun{
 #' library(ggplot2)
 #' library(gghighlight)
 #' library(dplyr)
@@ -68,6 +69,7 @@
 #'              y = ln_wages,
 #'              group = id)) + 
 #'   geom_line()
+#' }
 #'   
 increasing <- function(x, ...){
 

R/stratify-keys.R

@@ -30,7 +30,7 @@
 #'  facet_wrap(~.strata)
 #'  
 #'  # now facet along some feature
-#'  
+#'  \dontrun{
 #' library(dplyr)
 #'  wages %>%
 #' key_slope(ln_wages ~ xp) %>%
@@ -53,6 +53,7 @@
 #'              group = id)) + 
 #'   geom_line() + 
 #'   facet_wrap(~.strata)
+#' }
 stratify_keys <- function(.data, n_strata, along = NULL, fun = mean, ...){
   test_if_tsibble(.data)
   test_if_null(.data)

tests/testthat/test-feature-diff-summary.R

@@ -0,0 +1,26 @@
+context("feature diff summary")
+
+heights_diff_summary <- heights %>%
+  features(year, feat_diff_summary)
+
+test_that("feat_diff_summary returns the right names", {
+  expect_equal(names(heights_diff_summary),
+               c("country", 
+                 "diff_var",
+                 "diff_sd",
+                 "diff_min",
+                 "diff_q25",
+                 "diff_mean",
+                 "diff_median",
+                 "diff_q75",
+                 "diff_max"))
+})
+
+test_that("feat_diff_summary returns the right dimensions", {
+  expect_equal(dim(heights_diff_summary),
+               c(144, 9))
+})
+
+test_that("feat_diff_summary returns all ids", {
+  expect_equal(dplyr::n_distinct(heights_diff_summary$country), 144)
+})

tests/testthat/test-sample-n-frac-multiple-keys.R

@@ -7,8 +7,6 @@ context("multiple keys supported")
 
 data("aus_retail")
 
-aus_retail
-
 # will fail
 test_that("multiple keys doesn't fail",{
   expect_equal(ncol(sample_n_keys(aus_retail, size = 10)), 5)
@@ -37,7 +35,6 @@ sample_n_keys_retail_nkeys <- sample_n_keys(aus_retail, size = 10)
 sample_frac_keys_retail_nkeys <- sample_frac_keys(aus_retail, size = 0.11)
 n_keys(sample_frac_keys_retail_nkeys)
 
-aus_retail
 
 test_that("correct number of columns returned",{
   expect_equal(ncol(sample_n_keys_retail_nkeys), ncol(aus_retail))