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)

Calculating summaries by group with NumPy arrays

We can accomplish much of what we did in the previous recipe with itertuples using NumPy arrays. We can also use NumPy arrays to get summary values for subsets of our data.

Getting ready

We will work again with the COVID-19 case daily data and the Brazil land temperature data.

How to do it…

We copy DataFrame values to a NumPy array. We then navigate over the array, calculating totals by group and checking for unexpected changes in values:

  1. Import pandas and numpy, and load the Covid and land temperature data:
    >>> import pandas as pd
    >>> import numpy as np
    >>> coviddaily = pd.read_csv("data/coviddaily720.csv", parse_dates=["casedate"])
    >>> ltbrazil = pd.read_csv("data/ltbrazil.csv")
  2. Create a list of locations:
    >>> loclist = coviddaily.location.unique().tolist()
  3. Use a NumPy array to calculate sums by location.

    Create a NumPy array...