Data Wrangling with Python

Oil does not come in its final form from the rig; it has to be refined. Similarly, data must be curated, massaged, and refined to be used in intelligent algorithms and consumer products. This is known as wrangling. Most data scientists spend the majority of their time data wrangling.

Data wrangling is generally done at the very first stage of a data science/analytics pipeline. After the data scientists identify useful data sources for solving the business problem (for instance, in-house database storage or internet or streaming sensor data), they then proceed to extract, clean, and format the necessary data from those sources.

Generally, the task of data wrangling involves the following steps:

This is an illustrative representation of the positioning and essential functional role of data wrangling in a typical data science pipeline:

Figure 1.1: Process of data wrangling

The process of data wrangling includes first finding the appropriate data that's necessary for the analysis. This data can be from one or multiple sources, such as tweets, bank transaction statements in a relational database, sensor data, and so on. This data needs to be cleaned. If there is missing data, we will either delete or substitute it, with the help of several techniques. If there are outliers, we need to first detect them and then handle them appropriately. If data is from multiple sources, we will have to perform join operations to combine it.

In an extremely rare situation, data wrangling may not be needed. For example, if the data that's necessary for a machine learning task is already stored in an acceptable format in an in-house database, then a simple SQL query may be enough to extract the data into a table, ready to be passed on to the modeling stage.

Lists are fundamental Python data structures that have continuous memory locations, can host different data types, and can be accessed by the index.

We will start with a list and list comprehension. We will generate a list of numbers, and then examine which ones among them are even. We will sort, reverse, and check for duplicates. We will also see how many different ways we can access the list elements, iterating over them and checking the membership of an element.

The following is an example of a simple list:

list_example = [51, 27, 34, 46, 90, 45, -19]

The following is also an example of a list:

list_example2 = [15, "Yellow car", True, 9.456, [12, "Hello"]]

As you can see, a list can contain any number of the allowed datatype, such as int, float, string, and Boolean, and a list can also be a mix of different data types (including nested lists).

If you are coming from a strongly typed language, such as C, C++, or Java, then this will probably be strange as you are not allowed to mix different kinds of data types in a single array in those languages. Lists are somewhat like arrays, in the sense that they are both based on continuous memory locations and can be accessed using indexes. But the power of Python lists come from the fact that they can host different data types and you are allowed to manipulate the data.

In the following exercise, we will be creating a list and then observing the different ways of accessing the elements:

We are going to examine various ways of generating a list:

We are going to iterate over a list and test whether a certain value exists in it:

We generated a list called list_1 in the previous exercise. We are going to sort it now:

In this exercise, we will be generating a list with random numbers:

There are many ways to get a list of unique numbers, and while you may be able to write a few lines of code using a for loop and another list (you should actually try doing it!), let's see how we can do this without a for loop and with a single line of code. This will bring us to the next data structure, sets.

In this activity, we will generate a list of random numbers and then generate another list from the first one, which only contains numbers that are divisible by three. Repeat the experiment three times. Then, we will calculate the average difference of length between the two lists.

These are the steps for completing this activity:

A set, mathematically speaking, is just a collection of well-defined distinct objects. Python gives us a straightforward way to deal with them using its set datatype.

With the last list that we generated, we are going to revisit the problem of getting rid of duplicates from it. We can achieve that with the following line of code:

list_12 = list(set(list_1))

If we print this, we will see that it only contains unique numbers. We used the set data type to turn the first list into a set, thus getting rid of all duplicate elements, and then we used the list function on it to turn it into a list from a set once more:

list_12

The output will be as follows:

Figure 1.12: Section of output for list_21

This is what a union between two sets looks like:

Figure 1.13: Venn diagram showing the union of two sets

This simply means take everything from both sets but take the common elements only once.

We can create this using the following code:

set1 = {"Apple", "Orange", "Banana"}
set2 = {"Pear", "Peach", "Mango", "Banana"}

To find the union of the two sets, the following instructions should be used:

set1 | set2

The output would be as follows:

{'Apple', 'Banana', 'Mango', 'Orange', 'Peach', 'Pear'}

Notice that the common element, Banana, appears only once in the resulting set. The common elements between two sets can be identified by obtaining the intersection of the two sets, as follows:

Figure 1.14: Venn diagram showing the intersection of two sets

We get the intersection of two sets in Python as follows:

set1 & set2

This will give us a set with only one element. The output is as follows:

{'Banana'}

You can create a null set by creating a set containing no elements. You can do this by using the following code:

null_set_1 = set({})
null_set_1

The output is as follows:

set()

However, to create a dictionary, use the following command:

null_set_2 = {}
null_set_2

The output is as follows:

{}

We are going to learn about this in detail in the next topic.

A dictionary is like a list, which means it is a collection of several elements. However, with the dictionary, it is a collection of key-value pairs, where the key can be anything that can be hashed. Generally, we use numbers or strings as keys.

To create a dictionary, use the following code:

dict_1 = {"key1": "value1", "key2": "value2"}
dict_1

The output is as follows:

{'key1': 'value1', 'key2': 'value2'}

This is also a valid dictionary:

dict_2 = {"key1": 1, "key2": ["list_element1", 34], "key3": "value3",
          "key4": {"subkey1": "v1"}, "key5": 4.5}
dict_2

The output is as follows:

{'key1': 1,
 'key2': ['list_element1', 34],
 'key3': 'value3',
 'key4': {'subkey1': 'v1'},
 'key5': 4.5}

The keys must be unique in a dictionary.

In this exercise, we are going to access and set values in a dictionary:

In this exercise, we are going to iterate over a dictionary:

We will use the fact that dictionary keys cannot be duplicated to generate the unique valued list:

In this exercise, we are going to delete a value from a dict:

In this final exercise on dict, we will go over a less used comprehension than the list one: dictionary comprehension. We will also examine two other ways to create a dict, which will be useful in the future.

A dictionary comprehension works exactly the same way as the list one, but we need to specify both the keys and values:

A tuple is another data type in Python. It is sequential in nature and similar to lists.

A tuple consists of values separated by commas, as follows:

tuple_1 = 24, 42, 2.3456, "Hello"

Notice that, unlike lists, we did not open and close square brackets here.

This is how we create an empty tuple:

tuple_1 = ()

And this is how we create a tuple with only one value:

tuple_1 = "Hello",

Notice the trailing comma here.

We can nest tuples, similar to list and dicts, as follows:

tuple_1 = "hello", "there"
tuple_12 = tuple_1, 45, "Sam"

One special thing about tuples is the fact that they are an immutable data type. So, once created, we cannot change their values. We can just access them, as follows:

tuple_1 = "Hello", "World!"
tuple_1[1] = "Universe!"

The last line of code will result in a TypeError as a tuple does not allow modification.

This makes the use case for tuples a bit different than lists, although they look and behave very similarly in a few aspects.

The term unpacking a tuple simply means to get the values contained in the tuple in different variables:

tuple_1 = "Hello", "World"
hello, world = tuple_1
print(hello)
print(world)

The output is as follows:

Hello
World

Of course, as soon as we do that, we can modify the values contained in those variables.

We have mainly seen two different types of data so far. One is represented by numbers; another is represented by textual data. Whereas numbers have their own tricks, which we will see later, it is time to look into textual data in a bit more detail.

In the final section of this section, we will learn about strings. Strings in Python are similar to any other programming language.

This is a string:

string1 = 'Hello World!' 

A string can also be declared in this manner:

string2 = "Hello World 2!"

You can use single quotes and double quotes to define a string.

Strings in Python behave similar to lists, apart from one big caveat. Strings are immutable, whereas lists are mutable data structures:

Just like lists, we can slice strings:

To find out the length of a string, we simply use the len function:

str_1 = "Hello World! I am learning data wrangling"
len(str_1)

The length of the string is 41. To convert a string's case, we can use the lower and upper methods:

str_1 = "A COMPLETE UPPER CASE STRING"
str_1.lower()
str_1.upper()

The output is as follows:

'A COMPLETE UPPER CASE STRING'

To search for a string within a string, we can use the find method:

str_1 = "A complicated string looks like this"
str_1.find("complicated")
str_1.find("hello")# This will return -1

The output is -1. Can you figure out whether the find method is case-sensitive or not? Also, what do you think the find method returns when it actually finds the string?

To replace one string with another, we have the replace method. Since we know that a string is an immutable data structure, replace actually returns a new string instead of replacing and returning the actual one:

str_1 = "A complicated string looks like this"
str_1.replace("complicated", "simple")

The output is as follows:

'A simple string looks like this'

You should look up string methods in the standard documentation of Python 3 to discover more about these methods.

These two string methods need separate introductions, as they enable you to convert a string into a list and vice versa:

With these, we are at the end of our second topic of this chapter. We now have the motivation to learn data wrangling and have a solid introduction to the fundamentals of data structures using Python. There is more to this topic, which will be covered in a future chapters.

We have designed an activity for you so that you can practice all the skills you just learned. This small activity should take around 30 to 45 minutes to finish.

This section will ensure that you have understood the various basic data structures and their manipulation. We will do that by going through an activity that has been designed specifically for this purpose.

In this activity, we will do the following:

These are the steps to guide you through solving this activity: