R/sample_from_posterior.R

In grizbayr: Bayesian Inference for A|B and Bandit Marketing Tests

#' Sample From Posterior
#'
#' Selects which function to use to sample from the posterior distribution
#'
#' @param input_df Dataframe containing option_name (str) and various other columns
#'     depending on the distribution type. See vignette for more details.
#' @param distribution String of the distribution name
#' @param priors Optional list of priors. Defaults will be use otherwise.
#' @param n_samples Optional integer value. Defaults to 50,000 samples.
#'
#' @return A tibble with 2 columns: option_name (chr) and samples (dbl) [long form data].
#' @export
#' @importFrom dplyr select mutate row_number %>%
#' @importFrom tidyr unnest
#' @importFrom purrr map
#' @importFrom tibble tibble
#' @importFrom rlang .data
#'
#' @examples
#' input_df <- tibble::tibble(
#'    option_name = c("A", "B"),
#'    sum_clicks = c(1000, 1000),
#'    sum_conversions = c(100, 120),
#'    sum_sessions = c(1000, 1000),
#'    sum_revenue = c(1000, 1500)
#' )
#' sample_from_posterior(input_df, "conversion_rate")
#' sample_from_posterior(input_df, "rev_per_session")
#'
sample_from_posterior <- function(input_df, distribution, priors = list(), n_samples = 5e4){
  samples_tibble <- switch(distribution,
   "conversion_rate" =        sample_conv_rate(input_df, priors, n_samples),
   "response_rate" =          sample_response_rate(input_df, priors, n_samples),
   "ctr" =                    sample_ctr(input_df, priors, n_samples),
   "rev_per_session" =        sample_rev_per_session(input_df, priors, n_samples),
   "multi_rev_per_session" =  sample_multi_rev_per_session(input_df, priors, n_samples),
   "cpa" =                    sample_cpa(input_df, priors, n_samples),
   "total_cm" =               sample_total_cm(input_df, priors, n_samples),
   "cm_per_click" =           sample_cm_per_click(input_df, priors, n_samples),
   "cpc" =                    sample_cpc(input_df, priors, n_samples),
   "session_duration" =       sample_session_duration(input_df, priors, n_samples),
   "page_views_per_session" = sample_page_views_per_session(input_df, priors, n_samples),
   stop(
     paste(distribution,
           "is an invalid distribution type. Select from one of the following:",
           paste(distribution_column_mapping$distribution_type, collapse = ", "))
     )
  )
  # Clean tibble into expected 2 dimensional output with added sample_id
  samples_tibble %>%
     dplyr::mutate(samples = purrr::map(.x = .data$samples,
                                        ~ tibble::tibble(samples = .x) %>%
                                           dplyr::mutate(sample_id = dplyr::row_number())
                                        )
                   ) %>%
     dplyr::select("option_name", "samples") %>%
     tidyr::unnest(cols = "samples")
}