Data Cleaning with tidyr

library(tidyr)
library(dplyr)

# Wide data: year columns contain values
wide <- tibble(
  country = c("USA", "UK", "Japan"),
  `2020` = c(100, 80, 90),
  `2021` = c(110, 85, 95),
  `2022` = c(120, 90, 100)
)

# Pivot to tidy (long) format
long <- wide |>
  pivot_longer(
    cols = `2020`:`2022`,
    names_to = "year",
    values_to = "value"
  )
# country  year   value
# USA      2020   100
# USA      2021   110
# ...

# Convert back to wide format
long |>
  pivot_wider(
    names_from = year,
    values_from = value
  )

# Split one column into two
df <- tibble(date = c("2024-01-15", "2024-02-20", "2024-03-25"))
df |> separate(date, into = c("year", "month", "day"), sep = "-")

# Combine multiple columns into one
df2 <- tibble(first = c("Alice", "Bob"), last = c("Smith", "Jones"))
df2 |> unite("full_name", first, last, sep = " ")

df <- tibble(
  x = c(1, NA, 3, NA, 5),
  y = c(NA, 2, NA, 4, 5)
)

# Drop rows with any NA
df |> drop_na()

# Drop rows where specific column is NA
df |> drop_na(x)

# Replace NA with a value
df |> replace_na(list(x = 0, y = 0))

# Fill NA with previous/next value
df |> fill(x, .direction = "down")
df |> fill(x, .direction = "up")

# Complete cases (no NAs in any column)
df |> filter(complete.cases(df))

# Nest data by group
nested <- mtcars |>
  group_by(cyl) |>
  nest()
# Each group's data stored as a tibble in a list-column

# Fit a model to each group
library(purrr)
models <- nested |>
  mutate(model = map(data, \(df) lm(mpg ~ wt, data = df)))

# Unnest back to flat data
nested |> unnest(data)