Book Image

The Pandas Workshop

By : Blaine Bateman, Saikat Basak, Thomas V. Joseph, William So
5 (1)
Book Image

The Pandas Workshop

5 (1)
By: Blaine Bateman, Saikat Basak, Thomas V. Joseph, William So

Overview of this book

The Pandas Workshop will teach you how to be more productive with data and generate real business insights to inform your decision-making. You will be guided through real-world data science problems and shown how to apply key techniques in the context of realistic examples and exercises. Engaging activities will then challenge you to apply your new skills in a way that prepares you for real data science projects. You’ll see how experienced data scientists tackle a wide range of problems using data analysis with pandas. Unlike other Python books, which focus on theory and spend too long on dry, technical explanations, this workshop is designed to quickly get you to write clean code and build your understanding through hands-on practice. As you work through this Python pandas book, you’ll tackle various real-world scenarios, such as using an air quality dataset to understand the pattern of nitrogen dioxide emissions in a city, as well as analyzing transportation data to improve bus transportation services. By the end of this data analytics book, you’ll have the knowledge, skills, and confidence you need to solve your own challenging data science problems with pandas.
Table of Contents (21 chapters)
1
Part 1 – Introduction to pandas
6
Part 2 – Working with Data
11
Part 3 – Data Modeling
15
Part 4 – Additional Use Cases for pandas

Summarizing data

Summarizing data is one of the most important tasks in data analysis, as this is the step where a data analyst will convert a large amount of data into a few main aggregates that represent a summary of the data. First, you will learn about the basics of data aggregation with pandas. Then, we will move on to a more advanced topic with pivot tables.

Grouping and aggregation

In general, datasets are made of a single observation per row, which means that you can end up with datasets comprising millions of rows. Of course, deriving any data analysis on dozens of rows is not the same as millions of rows. In these situations, grouping/summarizing rows together based on common variables is a good solution.

Consider the following example. You are given a file containing the yearly sales of a number of stores, as follows:

Figure 7.76 – A sample DataFrame of sales

And you have been asked to summarize the sales for each store, which should...