Pivoting data

Data Manipulation in Julia

Katerina Zahradova

Instructor

Pivot tables

Way of summarizing data
Reorganization of columns and rows to make data more readable

Pivot table using unstack()

unstack structure

unstack() to pivot

# Using unstack to pivot
# unstack(DataFrame, :col_as_rows, :col_as_cols, :values, combine = f_to_aggregate)

unstack(penguins, :species, :sex, :body_mass_g, combine = median)

3×3 DataFrame
Row  species   MALE     FEMALE
     String15  Float64  Float64
_____________________________________
1    Adelie    4000.0   3400.0
2    Chinstrap 3950.0   3550.0
3    Gentoo    5500.0   4700.0

Pivoting on multiple columns as rows

# Pivot on more columns as rows 
unstack(penguins, [:species, :island], :sex, :body_mass_g, combine = median)

5×4 DataFrame
Row  species    island     MALE       FEMALE
     String15   String15   Float64?   Float64?
_____________________________________________________
1    Adelie     Torgersen  4000.0    3400.0
2    Adelie     Biscoe     4000.0    3375.0
3    Adelie     Dream      3987.5    3400.0
4    Chinstrap  Dream      3950.0    3550.0
5    Gentoo     Biscoe     5500.0    4700.0

Pivoting on multiple columns elsewhere

# Using multiple columns as columns
unstack(penguins, :sex, [:species, :island], :body_mass_g, combine = sum)

MethodError: no method matching 
unstack(::DataFrame, ::Symbol, ::Vector{Symbol}, ::Symbol; combine = sum)

Missing values

# Missing values for certain combinations
unstack(penguins_missing, :species, :sex, :body_mass_g, combine = median)

3×3 DataFrame
Row  species   MALE     FEMALE
     String15  Float64  Float64?
_____________________________________
1    Adelie    4000.0   3400.0
2    Chinstrap 3950.0   3550.0
3    Gentoo    5500.0   missing

Replacing missing values

# Missing values for certain combinations
unstack(penguins_missing, :species, :sex, :body_mass_g, combine = median, fill = -1)

3×3 DataFrame
Row  species   MALE     FEMALE
     String15  Float64  Float64
_____________________________________
1    Adelie    4000.0    3400.0
2    Chinstrap 3950.0    3550.0
3    Gentoo    5500.0    -1

Pivot tables are DataFrames

# Saving the pivot table
pivot_penguins = unstack(penguins_missing, :species, :sex, :body_mass_g, combine = median)

# Select only female penguins
select(pivot_penguins, :species, :FEMALE)

3×2 DataFrame
Row  species    FEMALE
     String15   Float64?
_______________________________
1    Adelie      3400.0
...

Let's practice!

Data Manipulation in Julia