Book Image

R Data Science Essentials

Book Image

R Data Science Essentials

Overview of this book

With organizations increasingly embedding data science across their enterprise and with management becoming more data-driven it is an urgent requirement for analysts and managers to understand the key concept of data science. The data science concepts discussed in this book will help you make key decisions and solve the complex problems you will inevitably face in this new world. R Data Science Essentials will introduce you to various important concepts in the field of data science using R. We start by reading data from multiple sources, then move on to processing the data, extracting hidden patterns, building predictive and forecasting models, building a recommendation engine, and communicating to the user through stunning visualizations and dashboards. By the end of this book, you will have an understanding of some very important techniques in data science, be able to implement them using R, understand and interpret the outcomes, and know how they helps businesses make a decision.

Related Content you might be interested in

Current Title:

Small book image
R Data Science Essentials
Jan, 2016 | 154 pages
No titles found
Table of Contents (15 chapters)
R Data Science Essentials
R Data Science Essentials
Credits
Credits
About the Authors
About the Authors
About the Reviewers
About the Reviewers
www.PacktPub.com
www.PacktPub.com
Preface
Preface
Free Chapter
1
Getting Started with R
Getting Started with R
Reading data from different sources
Reading data from a database
Data types in R
Data preprocessing techniques
Performing data operations
Control structures in R
Bringing data to a usable format
Summary
2
Exploratory Data Analysis
Exploratory Data Analysis
The Titanic dataset
Descriptive statistics
Inferential statistics
Univariate analysis
Bivariate analysis
Multivariate analysis
Summary
3
Pattern Discovery
Pattern Discovery
Transactional datasets
Apriori analysis
Support, confidence, and lift
Generating filtering rules
Plotting
Sequential dataset
Apriori sequence analysis
Understanding the results
Business cases
Summary
4
Segmentation Using Clustering
Segmentation Using Clustering
Datasets
Centroid-based clustering and an ideal number of clusters
Implementation using K-means
Visualizing the clusters
Connectivity-based clustering
Visualizing the connectivity
Business use cases
Summary
5
Developing Regression Models
Developing Regression Models
Datasets
Sampling the dataset
Logistic regression
Evaluating logistic regression
Linear regression
Evaluating linear regression
Methods to improve the accuracy
Ensemble models
Summary
6
Time Series Forecasting
Time Series Forecasting
Datasets
Extracting patterns
Forecasting using ARIMA
Forecasting using Holt-Winters
Methods to improve accuracy
Summary
7
Recommendation Engine
Recommendation Engine
Dataset and transformation
Recommendations using user-based CF
Recommendations using item-based CF
Challenges and enhancements
Summary
8
Communicating Data Analysis
Communicating Data Analysis
Dataset
Plotting using the googleVis package
Creating an interactive dashboard using Shiny
Summary
Index
Index

Datasets

In this chapter, we will use the temperature dataset of Massachusetts; this public dataset was downloaded from http://cdiac.ornl.gov/. This dataset holds the maximum temperature recorded at Massachusetts on a daily basis from 1980 to 2010. The temperature in this dataset is rounded off to an integer and the missing values are represented as NA. We will use this dataset to learn about the techniques involved in the forecasting algorithm.

Note

Note that changes in terms of representation of the data have been made to make the dataset more R-friendly in terms of reading and computing.

Let's have a look at the dataset by reading the dataset to the R environment:

# reading the dataset
data <- read.csv("Data/msdata.csv")
head(data, 10)

The output of the preceding code is as follows:

The preceding dataset needs some modifications, such as the date format has to be changed to one that will be supported for the time series analysis and the missing values in the dataset have to be replaced...

Previous Section
End of Section 2
Next Section