pandas DataFrames

A pandas DataFrame is the second type of class that is capable of handling data.

Similar to a spreadsheet, a DataFrame can be visualzed as having multiple columns and rows associated with the data inside. The data within can be of any type.

A DataFrame can also be considered a collection or assortment of Series. Similar to a Series there are multiple ways that a DataFrame can be created:

1. Using a multi-dimensional list, dictionary, or tuple
2. Combining or joining multiple Series together
3. From a pre-existing CSV file

Column values within a DataFrame are referred to as a Series. Below is an example of how multiple Series might be used to build a DataFrame

The image below provides another visual of the general DataFrame structure. A DataFrame is similar to a Python dictionary in that the column names are like keys and the values are the data for that column.

Creating a DataFrame

Let’s dive in to some different ways you can create a DataFrame.

Using a Multi-Dimensional List

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

# import pandas
import pandas as pd

# Create a pandas DataFrame by providing a list of lists
movie_list_of_lists = pd.DataFrame([["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],["Marley & Me", "Two Weeks Notice", "The Guardian", "Bridesmaids"]])

# Create a pandas DataFrame from a pre-existing list of lists
movies_dataframe_data = [["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],["Marley & Me", "The Proposal", "The Guardian", "Bridesmaids"]]

dataframe_from_existing_list = pd.DataFrame(movies_dataframe_data)

The above code block accomplishes the following:

1. imports pandas.
2. Creates a pandas DataFrame called movie_list_of_lists by providing a list of lists as a parameter into the .DataFrame() function..
3. Creates a pandas DataFrame called dataframe_from_existing_list by using the already existing list movies_dataframe_data and passing it in as a parameter to the .DataFrame() function.

Using a Dictionary

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

# import pandas
import pandas as pd

# Create a pandas DataFrame by providing a dictionary
movie_dictionary_dataframe = pd.DataFrame(movies = {'Name': ["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],'Release': [2014, 2005, 2010, 2003]})

# Create a pandas DataFrame from a pre-existing dictionary
movies = {'Name': ["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],'Release': [2014, 2005, 2010, 2003]}

dataframe_from_movies_dictionary = pd.DataFrame(movies)

The above code block accomplishes the following:

1. imports pandas.
2. Creates a pandas DataFrame called movie_dictionary_dataframe by providing a dictionary as a parameter to the .DataFrame() function.
3. Creates a pandas DataFrame called dataframe_from_movies_dictionary by using the already existing dictionary movies and passing it in as a parameter to the .DataFrame() function.

Using a Tuple

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

# import pandas
import pandas as pd

# Create a pandas DataFrame by providing a tuple
movies_tuple_dataframe = pd.DataFrame(("Interstellar", "Pride and Prejudice"), ("Inception", "Barbie"))

# Create a pandas DataFrame by providing a pre-existing tuple
movies_data = (("Interstellar", "Pride and Prejudice"), ("Inception", "Barbie"))

dataframe_from_existing_tuple = pd.DataFrame(movies_data)

The above code block accomplishes the following:

1. imports pandas.
2. Creates a pandas DataFrame called movies_tuple_dataframe by providing a tuple as a parameter to the .DataFrame() function.
3. Creates a pandas DataFrame called dataframe_from_existing_tuple by using an already existing tuple movies_data and passing it in as a parameter to the .DataFrame() function.

Creating a DataFrame from Series

In the following example we will create a DataFrame from two Series using pandas and the .concat() function included with the pandas library.

1
2
3
4
5
6
7
8
9

import pandas as pd

movies = pd.Series(["Interstellar", "Pride and Prejudice", "Inception", "Barbie"], index=['1', '2', '3', '4'], name = 'movies')

genres = pd.Series(["Science Fiction", "Novel", "Science Fiction", "Comedy"], index=['1', '2', '3', '4'], name='genres')

# Using the two series above we can concatenate the two together in order to create a DataFrame.

movies_genres_dataframe = pd.concat([movies, genres], axis=1) # axis 1 specifies that the operations will be performed down each column

                movies           genres
1         Interstellar  Science Fiction
2  Pride and Prejudice            Novel
3            Inception  Science Fiction
4               Barbie           Comedy

Column Data

Suppose you want to view data from one particular column or compare specific columns to one another. You can do so by using the column labels to pull them aside. Let’s take a look at how to do so using the same dictionary we created above.

# import pandas
import pandas as pd

movies = {'Name': ["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],'Release': [2014, 2005, 2010, 2003]}

movies_dataframe = pd.DataFrame(movies)

movie_names = movies_dataframe["Name"]

The above example accomplishes the following:

1. Imports pandas
2. Creates a dictionary called movies with the columns Name and Release.
3. Creates a DataFrame from the movies dictionary
4. A new variable called movie_names is created to store the values within the Name column of the movies_dataframe.

Multiple Column Data

Now that you have seen how to pull aside a single column’s data let’s take a look at how to grab multiple columns and store them inside of a variable.

# import pandas
import pandas as pd

movies = {'Name': ["Interstellar", "Pride and Prejudice", "Inception", "Barbie"],'Release': [2014, 2005, 2010, 2023], 'Genre': ["Science Fiction", "Novel", "Science Fiction", "Comedy"]}

movies_dataframe = pd.DataFrame(movies)

# Pull aside the Name and Genre columns from the movies_dataframe
movie_names_and_genres = movies_dataframe[["Name", "Genre"]]

Check Your Understanding