Reshaping Data

We’ve already covered how to sort or group by values in a dataframe. We can also perform more drastic data manipulations and transformations using pandas.

A digest of these operations, courtesty of Pandas documentation:

• pandas.pivot and pandas.pivot_table: Group unique values within one or more discrete categories.

• DataFrame.stack and DataFrame.unstack: Pivot a column or row level to the opposite axis respectively.

• pandas.melt and pandas.wide_to_long: Unpivot a wide DataFrame to a long format.

• pandas.get_dummies and pandas.from_dummies: Conversions with indicator variables.

• Series.explode: Convert a column of list-like values to individual rows.

• pandas.crosstab: Calculate a cross-tabulation of multiple 1 dimensional factor arrays.

• pandas.cut: Transform continuous variables to discrete, categorical values

• pandas.factorize: Encode 1 dimensional variables into integer labels.

We’ll cover some of these workflows in greater depth.

Hierarchical Indexing, or MultiIndex

As part of these workflows, we’ll see the different ways Pandas supports hierarchical indexing, sometimes called a multi-index.

Multiple rows or columns can be part of a DataFrame’s indices.

For more detailed information:

• Python Data Science Handbook

• Pandas documentation

This Section’s Contents

• Pivot
• Pivot Table
• Melt
• Transpose
• Explode