How to Reshape a Data Frame Using stack() and unstack() Functions in Pandas

1. the stack() and unstack() Functions in Pandas
2. Using unstack() Function to Alter Our Data Frame
3. Using unstack() Function to Alter Our Data Frame

Pandas is an advanced data analysis tool or a package extension in Python. It is highly recommended to use Pandas when we have data in a SQL table, a spreadsheet or heterogenous columns.

This article explores the basic concept of the stack and unstacking in Pandas. Stacking and unstacking are used in Pandas widely to alter the shape of the data frame under consideration.

Let us see this method in action. First, we will create a dummy data frame, dates_data, along with a few rows.

import pandas as pd

index = pd.date_range("2013-1-1", periods=100, freq="30Min")
dates_data = pd.DataFrame(data=list(range(100)), columns=["value"], index=index)
dates_data["value2"] = "Alpha"
dates_data["value2"].loc[0:10] = "Beta"

The above code block creates a data frame dates_data with dates and two columns named value and value2. Viewing the entries in the data, we use the following code:

print(dates_data)

Output:

                     value value2
2013-01-01 00:00:00      0   Beta
2013-01-01 00:30:00      1   Beta
2013-01-01 01:00:00      2   Beta
2013-01-01 01:30:00      3   Beta
2013-01-01 02:00:00      4   Beta
...                    ...    ...
2013-01-02 23:30:00     95  Alpha
2013-01-03 00:00:00     96  Alpha
2013-01-03 00:30:00     97  Alpha
2013-01-03 01:00:00     98  Alpha
2013-01-03 01:30:00     99  Alpha

As we can see, we have 100 different entries with time set up equally after intervals of 30 minutes each.

Moreover, two additional columns named value and value2 are created where we have some values set as numbers and others as either Alpha or Beta.

the stack() and unstack() Functions in Pandas

We can alter our data frame named dates_data with the help of two functions named stack() and unstack() in Pandas. This function can help us change the orientation of the data frame such that the rows become columns and the columns become rows accordingly.

We will try to alter value and value2 in our data frame as the rows and the values in those as the entries in our rows.

Using unstack() Function to Alter Our Data Frame

Command:

dates_data = dates_data.unstack()
print(dates_data)

Output:

value   2013-01-01 00:00:00        0
        2013-01-01 00:30:00        1
        2013-01-01 01:00:00        2
        2013-01-01 01:30:00        3
        2013-01-01 02:00:00        4
                               ...
value2  2013-01-02 23:30:00    Alpha
        2013-01-03 00:00:00    Alpha
        2013-01-03 00:30:00    Alpha
        2013-01-03 01:00:00    Alpha
        2013-01-03 01:30:00    Alpha
Length: 200, dtype: object

Now, we have successfully altered our data such that we now have our columns as row entries in our data.

Using unstack() Function to Alter Our Data Frame

Command:

dates_data = dates_data.stack()
print(dates_data)

Output:

2013-01-01 00:00:00  value         0
                     value2     Beta
2013-01-01 00:30:00  value         1
                     value2     Beta
2013-01-01 01:00:00  value         2
                               ...
2013-01-03 00:30:00  value2    Alpha
2013-01-03 01:00:00  value        98
                     value2    Alpha
2013-01-03 01:30:00  value        99
                     value2    Alpha
Length: 200, dtype: object

The column values are now stacked as rows in our data frame.

Therefore, with the help of the unstacking technique in Pandas, we can efficiently filter data based on our requirement as and when needed and convert the look of our data frame that might be necessary to visualize the data in a better way.