Created
October 24, 2022 11:47
-
-
Save nathancday/193138e156129d2535ea6682adf20972 to your computer and use it in GitHub Desktop.
Code snippet for demo and graphic in Probability-Proportional-To-Size-Sampling blogpost
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#' --- | |
#' title: Code snippet for demo and graphic in Probability-Proportional-To-Size-Sampling blogpost | |
#' author: nathancday--at-Github-- | |
#' date: 2022-10-24 | |
#' --- | |
# Libs -------------------------------------------------------------------- | |
library(cowplot) | |
library(tidyverse) | |
# Helpers ----------------------------------------------------------------- | |
sample_and_plot = function(x, weight = FALSE) { | |
if (!weight) { | |
sample_dat = sample_n(dat, 50) | |
plt_title = paste0("SRS #", x) | |
} else { | |
# this is the magic ... weight=frewg | |
sample_dat = sample_n(dat, 50, weight = freq) | |
plt_title = paste0("PPTSS #", x) | |
} | |
sample_pct_lbl = scales::percent(sum(sample_dat$freq) / sum(dat$freq)) | |
ggplot(dat, aes(query, freq)) + | |
geom_col(fill = "white") + | |
geom_col(data = sample_dat) + | |
annotate("text", x = 400, y = 7000, label = paste0("Sample covers: ", sample_pct_lbl)) + | |
scale_x_discrete(labels = NULL) + | |
labs(title = plt_title, | |
y = NULL, x = NULL) | |
} | |
# Ingest ----------------------------------------------------------------- | |
# CSV data from AI Powered Search by Grainger. Turnbull, Irwin | |
# https://github.com/o19s/visualizing-signals#step-by-step-setup | |
dat = read_csv("weighted_sampling/signals.csv") %>% | |
filter(type == "query") %>% | |
mutate(query = tolower(target)) %>% | |
count(query, sort = TRUE, name = "freq") %>% | |
slice_head(n = 500) %>% | |
mutate(query = fct_inorder(query)) | |
# Viz --------------------------------------------------------------------- | |
ylab = "Traffic frequency" | |
xlab = "Individual queries\nMost frequent ---> Least frequent" | |
p_all = ggplot(dat, aes(query, freq)) + | |
geom_col() + | |
scale_x_discrete(labels = NULL) + | |
labs(title = "All queries sorted by traffic frequency", | |
y = ylab, x = xlab) | |
p_uni_li = map(1:5, ~ sample_and_plot(.x)) | |
p_uni = cowplot::plot_grid(plotlist = p_uni_li, ncol = 1) | |
p_wt_li = map(1:5, ~ sample_and_plot(.x, weight = TRUE)) | |
p_wt = cowplot::plot_grid(plotlist = p_wt_li, ncol = 1) | |
p_samples = cowplot::plot_grid(p_uni, p_wt, nrow = 1) | |
cowplot::plot_grid(p_all, p_samples, ncol = 1, rel_heights = c(0.33, .66)) | |
# Output ------------------------------------------------------------------ | |
ggsave("sampling.png", height = 7, width = 8) # saves last plot by default | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment