Data Science Programming In Python

Introduction - Data Science Programming in Python

The aim of this book is to show how to apply data analysis principles to a practical use case scenario using Python as the data analysis language. We’ll go on this journey by looking at the the data workflow from munging to grouping data to visualizing and also include some time-series analysis as well. The format includes asking questions of the data and showing the programming steps needed to answer the question. By the end of reading this book, you will be able to apply these techniques to your own data.

About the book

This book is written in a literate programming style where text, code, and output are presented together . This will maximize your learning and understanding of code and the data analysis workflow. The book teaches the type of interactive coding and iterative analysis that is essential to be successful in data science programming.    

Coding Tips

In the code snippets, a backslash character (\) means that the same line of code is wrapped to the next line in the book. You do not need to type this character into an interpreter.

Use a REPL (en.wikipedia.org/wiki/Read–eval–print_loop) to have an interactive environment where you can write code and see the resulting output.

Try the methods you learn in this book on your own data to reinforce learning. Use a Python interpreter to code and your favorite editor to take notes.

Data Science Programming in Python - Data Munging

Background

Bay Area Bike Share commenced it’s pilot phase of operation in the San Francisco bay area in August 2013 with plans to expand. It is the first bike sharing scheme in California. As it is meant for short trips, the bikes should be returned to a dock in thirty minutes or less or an additional fee would be incurred according to the website. There are two types of memberships: customer and subscriber. A subscriber is an annual membership while a customer is defined as someone using either the twenty-four hour or three day passes. Currently(Sept 2014), it costs nine dollars for twenty-four hours, twenty-two dollars for three days, and eight-eight dollars for the year. Overtime fees are four dollars for an extra thirty minutes and seven dollars for each thirty minutes after that. Data on the first six months of operations were released as part of a data challenge. The data included three files for trip history, weather information, and dock availability. The merged data was used for the following analysis.

Data Munging and Carpentry

First, we’ll read in the data and inspect the data columns and datatypes and think about what questions we want to ask our data and what things are we interested in learning about the data. Be curious and empathetic in thinking about what the various stakeholders including the City, the customers, and other interested people would be interested in gleaning by keeping civic fiscal, civic, and social goals in mind. In addition to that, there will be