Diwali Sales Data

library(tidyverse) # ggplot, lubridate, dplyr, stringr, readr...
library(tidytext)
library(praise)
library(paletteer)
library(ggforce)
library(networkD3)

The Data

The data this week comes from sales data for a retail store during the Diwali festival period in India. The data is shared on Kaggle by Saad Haroon. It isn’t clear how the data were obtained, including the personal characteristics. ???

diwali <- read_csv("diwali_sales_data.csv") |>
  janitor::clean_names() |>
  mutate(marital_status = as.factor(marital_status))

diwali |>
  group_by(product_category) |>
  summarize(number = n()) |>
  arrange(desc(number))

# A tibble: 18 × 2
   product_category      number
   <chr>                  <int>
 1 Clothing & Apparel      2655
 2 Food                    2493
 3 Electronics & Gadgets   2087
 4 Footwear & Shoes        1064
 5 Household items          520
 6 Beauty                   422
 7 Games & Toys             386
 8 Sports Products          356
 9 Furniture                353
10 Pet Care                 212
11 Office                   113
12 Stationery               112
13 Books                    103
14 Auto                     100
15 Decor                     96
16 Veterinary                81
17 Tupperware                72
18 Hand & Power Tools        26

diwali |>
  group_by(age_group,gender, marital_status) |>
  summarize(ave_amount = mean(amount, na.rm = TRUE)) |>
  ggplot(aes(x = age_group)) + 
  geom_bar() + 
  facet_wrap(gender ~ marital_status)

diwali |>
  filter(age < 20) |>
  ggplot(aes(x = age, y = amount, color = occupation)) +
  geom_point()

diwali |>
  ggplot(aes(x = product_category, y = amount, color = marital_status)) +
  geom_boxplot(position = position_dodge(width = 0.75)) + 
  theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1)) +
  ylim(c(10000,25000))

diwali |>
  group_by(state) |>
  summarize(number = n()) |>
  arrange(desc(number))

# A tibble: 16 × 2
   state            number
   <chr>             <int>
 1 Uttar Pradesh      1946
 2 Maharashtra        1526
 3 Karnataka          1305
 4 Delhi              1107
 5 Madhya Pradesh      923
 6 Andhra Pradesh      812
 7 Himachal Pradesh    608
 8 Kerala              453
 9 Haryana             452
10 Bihar               434
11 Gujarat             429
12 Jharkhand           380
13 Uttarakhand         320
14 Rajasthan           231
15 Punjab              200
16 Telangana           125

Making a Sankey plot to identify whether there are occupation trends across different regions. I’m slightly confused about the relationship between zone and state as I would have thought the state would be entirely contained in a particular zone. Regardless, the occupations do not seem to differ across the different states or regions.

diwali_table <- diwali |>
  filter(occupation %in% c("IT Sector", "Healthcare", "Aviation", "Banking",
                           "Govt", "Hospitality", "Media", "Automobile")) |>
  filter(state %in% c("Uttar Pradesh", "Maharashtra", "Karnataka", "Delhi",
                      "Madhya Pradesh", "Andhra Pradesh", "Himachal Pradesh")) |>
  select(zone, state, occupation) |>
  table()

data <- reshape2::melt(diwali_table)
data <- gather_set_data(data, 1:3) |>
  mutate(x = as.factor(x))

data |>
  mutate(x = fct_recode(x, "zone" = "1", "state" = "2", "occupation" = "3")) |>
ggplot(aes(x, id = id, split = y, value = value)) +
  geom_parallel_sets(aes(color = zone, fill = zone), alpha = 0.3, sep = 0.8) +
  geom_parallel_sets_axes(color = "lightgrey", fill = "white",
                          sep = 0.8, 
                          axis.width = c(rep(0.1, 4), rep(0.6, 7), rep(0.1, 8))) +
  geom_parallel_sets_labels(angle = c(rep(0, 4), rep(0, 15)), sep = 0.8, 
                            position = position_nudge(c(rep(-.3, 4), rep(0.01, 7),
                                                        rep(.3, 8))),

                            axis.width = c(rep(0.1, 4), rep(0.6, 7), rep(0.1, 8))) +
  theme(legend.position = "none", axis.title.y = element_blank(), 
        axis.text.y = element_blank(),
        axis.ticks.y = element_blank()) +
  xlab("") +
  labs(title = "Zone to State to Occupation",
       caption = "Data credit: @Saad Haroon")

Sankey plot showing how zone feeds into state which feeds into occupation.

praise()

[1] "You are cool!"