Laptop Prices with Tidyverse

# tidyverse loads the tools used to import, clean, reshape, summarize, and graph the data.
# knitr is used to display the final tables in a readable format.
library(tidyverse)
library(knitr)

# Read the laptop data stored in the data folder.
# show_col_types = FALSE keeps the rendered report from printing a long import message.
laptops_raw <- read_csv("data/laptops.csv", show_col_types = FALSE)

# Display the first six rows, but only use columns that fit on the page.
laptops_raw |>
  select(brand, model, device_category, price, ram_gb, storage_gb) |>
  slice_head(n = 6) |>
  kable(
    col.names = c("Brand", "Model", "Category", "Price (INR)", "RAM (GB)", "Storage (GB)"),
    caption = "First six laptop records"
  )

# Put the basic checks into one small table.
# These checks confirm the size of the file and show whether any values are missing.
data_check <- tibble(
  check = c("Rows", "Columns", "Missing cells", "Duplicate rows"),
  result = c(
    nrow(laptops_raw),
    ncol(laptops_raw),
    sum(is.na(laptops_raw)),
    sum(duplicated(laptops_raw))
  )
)

data_check |>
  kable(
    col.names = c("Check", "Result"),
    caption = "Basic data-quality checks"
  )

# First, count the missing cells in each column.
missing_by_column <- laptops_raw |>
  summarise(across(everything(), ~ sum(is.na(.x)))) |>
  # Turn the wide result into two columns: column name and missing count.
  pivot_longer(
    cols = everything(),
    names_to = "column",
    values_to = "missing_values"
  ) |>
  # Keep only the columns that have at least one missing value.
  filter(missing_values > 0) |>
  arrange(desc(missing_values))

missing_by_column |>
  kable(
    col.names = c("Column", "Missing values"),
    caption = "Columns that contain missing values"
  )

# Keep only the six columns used later in the report.
laptops <- laptops_raw |>
  select(
    brand, model, device_category, price, ram_gb, storage_gb
  ) |>
  mutate(
    # Change brand names to title case and correct the HP label.
    brand = str_to_title(brand),
    brand = if_else(brand == "Hp", "HP", brand),
    # Mark the laptops that meet the RAM and storage example.
    replacement_example = if_else(
      ram_gb >= 16 & storage_gb >= 512,
      "Meets 16 GB + 512 GB example",
      "Does not meet example"
    )
  )

# Count both groups and calculate the percentage of all laptops in each group.
standard_summary <- laptops |>
  count(replacement_example, name = "laptops") |>
  mutate(percent = scales::percent(laptops / sum(laptops), accuracy = 0.1))

standard_summary |>
  kable(
    col.names = c("Example group", "Laptops", "Percent"),
    caption = "Laptop records by the example replacement standard"
  )

# Group the records by laptop category.
# The median is used because a few very expensive laptops can pull the average upward.
category_summary <- laptops |>
  group_by(device_category) |>
  summarise(
    laptops = n(),
    median_price_inr = median(price),
    .groups = "drop"
  ) |>
  arrange(desc(laptops))

category_summary |>
  mutate(median_price_inr = scales::comma(median_price_inr)) |>
  kable(
    col.names = c("Laptop category", "Number of laptops", "Median price (INR)"),
    caption = "Laptop count and median price by category"
  )

# Make a dot chart from the same category_summary table shown above.
# The line helps connect each category to its value, and the exact median is printed.
ggplot(
  category_summary,
  aes(x = median_price_inr, y = reorder(device_category, median_price_inr))
) +
  geom_point(color = "#009E73", size = 4) +
  geom_text(
    aes(label = scales::comma(median_price_inr)),
    hjust = -0.08,
    size = 3.5
  ) +
  scale_x_continuous(
    labels = scales::label_number(big.mark = ","),
    expand = expansion(mult = c(0, 0.18))
  ) +
  labs(
    title = "Median laptop price by category",
    subtitle = "Prices are listed in Indian rupees",
    x = "Median price (INR)",
    y = NULL
  ) +
  theme_minimal(base_size = 11) +
  theme(panel.grid.major.y = element_blank())