Book Image

Python Data Cleaning Cookbook

By : Michael Walker
Book Image

Python Data Cleaning Cookbook

By: Michael Walker

Overview of this book

Getting clean data to reveal insights is essential, as directly jumping into data analysis without proper data cleaning may lead to incorrect results. This book shows you tools and techniques that you can apply to clean and handle data with Python. You'll begin by getting familiar with the shape of data by using practices that can be deployed routinely with most data sources. Then, the book teaches you how to manipulate data to get it into a useful form. You'll also learn how to filter and summarize data to gain insights and better understand what makes sense and what does not, along with discovering how to operate on data to address the issues you've identified. Moving on, you'll perform key tasks, such as handling missing values, validating errors, removing duplicate data, monitoring high volumes of data, and handling outliers and invalid dates. Next, you'll cover recipes on using supervised learning and Naive Bayes analysis to identify unexpected values and classification errors, and generate visualizations for exploratory data analysis (EDA) to visualize unexpected values. Finally, you'll build functions and classes that you can reuse without modification when you have new data. By the end of this Python book, you'll be equipped with all the key skills that you need to clean data and diagnose problems within it.
Table of Contents (12 chapters)

Identifying and cleaning missing data

We have already explored some strategies for identifying and cleaning missing values, particularly in Chapter 1, Anticipating Data Cleaning Issues when Importing Tabular Data into pandas. We will polish up on those skills in this recipe. We will do this by exploring a full range of strategies for handling missing data, including using DataFrame means and group means, as well as forward filling with nearby values. In the next recipe, we impute values using k-nearest neighbor.

Getting ready

We will continue working with the National Longitudinal Survey data in this recipe.

How to do it…

In this recipe, we will check key demographic and school record columns for missing values. We'll then use several strategies to impute values for missing data: assigning the overall mean for that column, assigning a group mean, and assigning the value of the nearest preceding non-missing value. Let's get started:

  1. Import pandas...