Setup

library(tidyverse)
library(lubridate)
library(knitr)

Chunk 1: Read, sort, filter

datafile <- "https://raw.githubusercontent.com/sebastiansauer/sesa-blog/main/static/datasets/sales_2008-2011.csv"

d <- read_csv(datafile)

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   product = col_character(),
##   country = col_character(),
##   date = col_date(format = ""),
##   quantity = col_double(),
##   amount = col_double(),
##   card = col_character(),
##   Cust_ID = col_character()
## )

glimpse(d)

## Rows: 47
## Columns: 7
## $ product  <chr> "prod_4", "prod_3", "prod_3", "prod_3", "prod_3", "prod_3", …
## $ country  <chr> "unknown", "China", "China", "China", "USA", "Brazil", "USA"…
## $ date     <date> 2008-12-12, 2009-04-10, 2009-04-10, 2009-05-10, 2009-05-20,…
## $ quantity <dbl> 1, 2, 2, 2, 20, 15, 2, 2, 20, 15, 15, 1, 1, 20, 1, 1, 25, 2,…
## $ amount   <dbl> 3, 160, 160, 160, 1600, 1200, 70, 70, 1600, 600, 600, 35, 35…
## $ card     <chr> NA, "N", "Y", NA, NA, NA, "Y", NA, NA, NA, "N", "Y", "Y", NA…
## $ Cust_ID  <chr> "Cust_8", "Cust_2", "Cust_5", "Cust_2", "Cust_3", "Cust_7", …

The data is already recognized as date; no need for transformation.

d %>% 
  arrange(date) %>% 
  select(-card) -> d

Chunk 2: group and aggregate

d %>% 
  group_by(Cust_ID) %>% 
  summarise(amount_sum = sum(amount),
            country = first(country),
            start = min(date),
            end = max(date),
            duration = end - start) -> d_summarized

## `summarise()` ungrouping output (override with `.groups` argument)

Chunk 3: filter

d_summarized %>% 
  filter(amount_sum > 5000) %>% 
  filter(country == "USA") -> d_summarized_filtered1

d_summarized %>% 
  filter(between(amount_sum,1000, 2000)) %>% 
  filter(country == "Germany") -> d_summarized_filtered2