Categorical Data Anaylsis is concerned with the analysis of categorical responses measures, regardless of whether any accompanying explanatory variables are also categorical or are continuous.

• An important consideration is the appropriate analysis of categorical variables is their scale of measurement

• Another consideration is the sampling framework that produced the data

  • determining the possible analyses and possible inferences

• Target populations generally assumed for each type of analysis

  • and what types of inferences you are able to make them

CategoricalArrays.jl

julia Pkg designed for either unorderded/nominal or ordered/ordinal categorical data

`CategoricalArray{T}` `CategoricalValue{T}`

• it maintains a pool of the levels which can appear in the given data
  • levels are stored in a specific order
  • for unordered arrays
    • this order is only used for pretty printing
  • for ordered arrays
    • it also allows comparing values using the `<` and `>` operators
      • comparison is then based on the ordering of levels stored in the array
    • for ordered `CategoricalValue` can also be compared with a value
      • that when converted is equal to one of the levels of this `CategoricalValue`

`ordered`, `ordered!` `levels`, `levels!` `droplevels!`

Categorical Data in DataFrames.jl

we have to deal with columns that take on a number of levels

we can represent the data more efficiently by replacing a naive encoding of every entry as a string with indices into a small pool of levels

• uses less memory!
• they can be grouped using `groupby`

PooledVector v.s. CategoricalVector

`PooledVector` is intended for cases where data compression is the only objective `CategoricalVector` is designed to additionally provide full support for working with categorical variables, both with unordered (nominal variables) and ordered categories (ordinal variables)