Pandas Series

Series in Pandas can be compared to one of the fundamental building blocks of the library that helps to manipulate and handle data. Essentially, it is a one-dimensional array with labels and indices.

What is the difference between a NumPy array and a Series?

Series in Pandas differ from a NumPy array because Series have labels on the elements, which NumPy arrays don't have. Thus you can access the Series elements with labels associated with them as well as the integer position of the element.

Parameters in Series:

Series consists of 4 main parameters:

1. data: Out of which the Series will be made

2. index: This will be used to label your data

3. dtype: Tells about the data type of the elements in the Series

4. copy: Input data is copied using this.

Let's use pandas Series:

Note: For a better understanding of the code, check the code here: colab.research.google

To begin, we import the pandas library and the NumPy library,

import pandas as pd

What does "import pandas as pd" signify?

Basically what this line of code does is, it imports the Pandas library and lets you call the library with the name pd.

Now let us create a Series:

After importing pandas you will be able to use pd.Series() function to create Series data structures. To do this we will pass an array of our choice into pd.Series().

import pandas as pd

studyTonightSeries= pd.Series([3,4,5,7,8,9])
print(studyTonightSeries)

0 3 1 4 2 5 3 7 4 8 5 9 dtype: int64

The column on the left-hand side of your output shows the index of each element present in the Series.

• We can create Series with NumPy and Arrays too.

• We first create a numpy array and then pass it into a pd.Series() function.

Below we have the code,

import pandas as pd
import numpy as np

st_ar = np.array(['s','t','u','d',’y’])
studyTonightSeries_ar= pd.Series(st_ar)
studyTonightSeries_ar

0 s 1 t 2 u 3 d 4 y dtype: object

As we can see our Series can be made up of any type of elements, strings, integers, float, etc.

Providing custom index in a Series:

Now let's see how we can set our own custom values for the index of pandas Series. Let's create a list of our index values,

ind = ['Row 1', 'Row 2', 'Row 3', 'Row 4', 'Row 5']

We can map this list onto our pandas Series object so that the list values can serve as an index to the Series object,

ind = ['Row 1', 'Row 2', 'Row 3', 'Row 4', 'Row 5']

studyTonight_arr = pd.Series([2,6,-3,1,7], index=ind)
print(studyTonight_arr)

Row 1 2 Row 2 6 Row 3 -3 Row 4 1 Row 5 7 dtype: int64

The index list is passed as a parameter into pd.Series() function.

Making Series from a dictionary in Python

If you have a dictionary in Python, you can turn it into a series. When you convert a python dictionary to pandas Series object, you will notice that the keys of the dictionary have become the index of the series:

studyTonight_dict = { 'Carrot': 12.9, 'Brinjal': 8.4, 'Gourd': 9.7 }

To convert the dictionary above, into a pandas Series, we need to pass the dictionary object into the Series() function as a parameter.

studyTonight_arr1 = pd.Series(studyTonight_dict)
print(studyTonight_arr1)

Carrot 12.9 Brinjal 8.4 Gourd 9.7 dtype: float64

Operations on Series:

Now let's see a few operations that we can perform on the series data structure in pandas library.

1. Get a Value from Series

To check the value corresponding to any row, we simply pass the name of the row as a parameter, as shown below.

import pandas as pd

ind = ['Row 1', 'Row 2', 'Row 3', 'Row 4', 'Row 5'] 
studyTonight_arr = pd.Series([2,6,-3,1], index=ind)

print(studyTonight_arr['Row 2'])

This command will return a value of 6.

We can also filter through a series by providing a condition instead of the index value. The resultant series will be filtered based on the condition which is provided. For example, if we want to get all the data elements stored in the series object with values greater than 2, we can do so like this,

import pandas as pd

ind = ['Row 1', 'Row 2', 'Row 3', 'Row 4', 'Row 5']
studyTonight_arr = pd.Series([2,6,-3,1], index=ind)

print(studyTonight_arr[studyTonight_arr > 2])

Row 2 6 Row 5 7 dtype: int64

As shown in the output, 6 and 7 are the only values present which are greater than 2.

2. Check if a value is present in Series

To check the presence of an item in a series we can use the in keyword. Using the in keyword will return a boolean value telling us if a particular item is present in the series or not. Let's take an example,

import pandas as pd

ind = ['Row 1', 'Row 2', 'Row 3', 'Row 4', 'Row 5']
studyTonight_arr = pd.Series([2,6,-3,1], index=ind)

'Row 3' in studyTonight_arr

The above statement will return True.

3. Mathematical operation on Series

Pandas series can be operated upon, mathematically. You can multiply, add, subtract and divide constants from an existing Series.

# to multiply all the values by 3
studyTonight_arr*3

Row 1 6 Row 2 18 Row 3 -9 Row 4 3 Row 5 21 dtype: int64

4. Demonstration of missing values:

To demonstrate the role of missing values in Series, we will first make a new list of items and add them to our Series.

vegies = ['Carrot', 'Brinjal', 'Peas', 'Gourd']

studyTonight_dict = { 'Carrot': 12.9, 'Brinjal': 8.4, 'Gourd': 9.7 }

studyTonight_arr2 = pd.Series(studyTonight_dict, index = vegies)

print(studyTonight_arr2)

In our newly built series, we have a new item “Peas” But as we can see our Series doesn’t have a value to correspond “Peas”. Therefore it is automatically represented in our System as NaN.

Carrot 12.9 Brinjal 8.4 Peas NaN Gourd 9.7 dtype: float64

NaN is the Pandas' method of representing missing values.

5. Adding two series:

You can perform arithmetic operations on 2 or more Series objects too. We can simply use the + operator to add two series objects.

studyTonight_dict = { 'Carrot': 12.9, 'Brinjal': 8.4, 'Gourd': 9.7 }
studyTonight_arr1 = pd.Series(studyTonight_dict)

vegies = ['Carrot', 'Brinjal', 'Peas', 'Gourd']
studyTonight_arr2 = pd.Series(studyTonight_dict, index = vegies)

print(studyTonight_arr1 + studyTonight_arr2)

Brinjal 16.8 Carrot 25.8 Gourd 19.4 Peas NaN dtype: float64

Adding two series objects will also add the respective values, but this is only possible if the datatype of the values store in the series being added is same. Also, the values are added only if the index values are same. If the index values are not the same then all the values will be together stored in the new series object.

dict_one = { 'Carrot': 12.9, 'Brinjal': 8.4, 'Gourd': 9.7 }
studyTonight_arr1 = pd.Series(dict_one)

dict_two = { 'Bread': 20.5, 'Eggs': 12.5, 'Milk': 21 }
studyTonight_arr2 = pd.Series(dict_two)

print(studyTonight_arr1 + studyTonight_arr2)

Bread NaN Brinjal NaN Carrot NaN Eggs NaN Gourd NaN Milk NaN dtype: float64

Notice the alphabetical order in which the index values got arranged. But all the values are changed to NaN.

6. Accessing a range of elements in a series:

The : operator in python lets us access a segment of lists etc and in this case our Series objects. Using it, we can access segments like the last 'n' elements, or the first 'n' elements or 'n' elements in between.

To get the first 2 elements from a Series, we use:

studyTonight_arr2[:2]

Carrot 12.9 Brinjal 8.4 dtype: float64

To get the last 2 elements from a Series, we use:

studyTonight_arr2[2:]

Peas NaN Gourd 9.7 dtype: float64

Therefore we can understand that the function essentially works in the way of [a:b] where a is the first element of our desired range and the b is the last element of our desired range. Using this we can take out a range of values from the middle too,

studyTonight_arr2[1:3]

Brinjal 8.4 Peas NaN dtype: float64

Conclusion:

In this tutorial, we covered the essential parts of the pandas Series object and also learned how to perform various functions over this data structure. The series data structure can be thought to be the second most important data structure in the Pandas library, therefore it is very important to get your basics cleared.