Book Image

Hands-On Data Science with R

By : Vitor Bianchi Lanzetta, Doug Ortiz, Nataraj Dasgupta, Ricardo Anjoleto Farias
Book Image

Hands-On Data Science with R

By: Vitor Bianchi Lanzetta, Doug Ortiz, Nataraj Dasgupta, Ricardo Anjoleto Farias

Overview of this book

R is the most widely used programming language, and when used in association with data science, this powerful combination will solve the complexities involved with unstructured datasets in the real world. This book covers the entire data science ecosystem for aspiring data scientists, right from zero to a level where you are confident enough to get hands-on with real-world data science problems. The book starts with an introduction to data science and introduces readers to popular R libraries for executing data science routine tasks. This book covers all the important processes in data science such as data gathering, cleaning data, and then uncovering patterns from it. You will explore algorithms such as machine learning algorithms, predictive analytical models, and finally deep learning algorithms. You will learn to run the most powerful visualization packages available in R so as to ensure that you can easily derive insights from your data. Towards the end, you will also learn how to integrate R with Spark and Hadoop and perform large-scale data analytics without much complexity.
Table of Contents (16 chapters)

Cloud computing

From my first year as an undergraduate student, there was this story a professor told us about top-notch computers taking weeks to fit a single linear regression model, back in her days as an undergrad student. I recall thinking, It's lucky that today's computers can run it in no time. This was naive though. As the history of computing shows, every time a faster horse is born, a more challenging track is built.

Computers have advanced a lot and so have the models—thankfully. In addition, our capacity to gather data has improved by teraflops. The first time I felt the need for cloud computing, I had designed a model so huge that I couldn't load it and the data simultaneously. It was either one or the other. Using a cloud service, this problem became manageable.

Using distributed computing solutions, such as Hadoop and Spark, is another way...