Tidy data

Tidy data are structured for use in R and satisfy three rules:

1. Each variable must have its own column
2. Each observation must have its own row
3. Each value must have its own cell.

(Grolemund and Wickham 2017:149)

There are two of tidy data formats: wide and long.

Tidy data

Source: Data Science with R, Garrett Grolemund

Why tidy data?

“Tidy datasets are all alike but every messy dataset is messy in its own way.”
– Hadley Wickham

Data

wide <- data.frame(Date = c("01-01-2017","02-01-2017","03-01-2017"),
                    Burglary = c(7,5,13),
                    Drugs = c(1,3,9),
                    Robbery = c(9,0,9),
                    Shoplifting = c(4,5,1))
wide

##         Date Burglary Drugs Robbery Shoplifting
## 1 01-01-2017        7     1       9           4
## 2 02-01-2017        5     3       0           5
## 3 03-01-2017       13     9       9           1

gather()

For when your columns are values, not variables (wide to long)

long <- gather(wide, category, count, -Date)
long

#          Date    category count
# 1  01-01-2017    Burglary     7
# 2  02-01-2017    Burglary     5
# 3  03-01-2017    Burglary    13
# 4  01-01-2017       Drugs     1
# 5  02-01-2017       Drugs     3
# 6  03-01-2017       Drugs     9
# 7  01-01-2017     Robbery     9
# 8  02-01-2017     Robbery     0
# 9  03-01-2017     Robbery     9
# 10 01-01-2017 Shoplifting     4
# 11 02-01-2017 Shoplifting     5
# 12 03-01-2017 Shoplifting     1

spread()

For when observations have multiple rows (long to wide)

wide <- spread(long, category, count)
wide

#         Date Burglary Drugs Robbery Shoplifting
# 1 01-01-2017        7     1       9           4
# 2 02-01-2017        5     3       0           5
# 3 03-01-2017       13     9       9           1

separate()

Split columns

split <- separate(wide, Date, c("Day", "Month", "Year"), sep = "-")
split

#   Day Month Year Burglary Drugs Robbery Shoplifting
# 1  01    01 2017        7     1       9           4
# 2  02    01 2017        5     3       0           5
# 3  03    01 2017       13     9       9           1

unite()

Join columns

united <- unite(split, Date, c(Day, Month, Year), sep = "-")
united

#         Date Burglary Drugs Robbery Shoplifting
# 1 01-01-2017        7     1       9           4
# 2 02-01-2017        5     3       0           5
# 3 03-01-2017       13     9       9           1