ggplot2 (part 1)

Loading...

From the course by Johns Hopkins University

Exploratory Data Analysis

4334 ratings

Johns Hopkins University

Exploratory Data Analysis

4334 ratings

Course 4 of 10 in the Specialization Data Science

This course covers the essential exploratory techniques for summarizing data. These techniques are typically applied before formal modeling commences and can help inform the development of more complex statistical models. Exploratory techniques are also important for eliminating or sharpening potential hypotheses about the world that can be addressed by the data. We will cover in detail the plotting systems in R as well as some of the basic principles of constructing data graphics. We will also cover some of the common multivariate statistical techniques used to visualize high-dimensional data.

From the lesson

Week 2

Welcome to Week 2 of Exploratory Data Analysis. This week covers some of the more advanced graphing systems available in R: the Lattice system and the ggplot2 system. While the base graphics system provides many important tools for visualizing data, it was part of the original R system and lacks many features that may be desirable in a plotting system, particularly when visualizing high dimensional data. The Lattice and ggplot2 systems also simplify the laying out of plots making it a much less tedious process.

ggplot2 (part 1)6:25

ggplot2 (part 2)13:53

ggplot2 (part 3)9:46

ggplot2 (part 4)10:37

ggplot2 (part 5)8:10

Meet the Instructors

Roger D. Peng, PhD
Associate Professor, Biostatistics
Bloomberg School of Public Health
Jeff Leek, PhD
Associate Professor, Biostatistics
Bloomberg School of Public Health
Brian Caffo, PhD
Professor, Biostatistics
Bloomberg School of Public Health

Hi, everyone, this lesson is going to be about ggplot2 or the ggplot2 package.

And we're going to, I'm going to talk about how to

do some basic plots using the ggplot2 package and what it's about.

And in this, in the next lecture I'll talk a little bit in more detail

about how it's designed and how you

can make extensions to various ggplot2 plotting functions.

So the first question very basic. You know, what is ggplot2?

Basically it's a package

in R that you can download from CRAN.

And and it implements what's called the

grammar of graphics, which is originally written by

Leland Wilkinson and it is described in a, in a book called the Grammar of Graphics.

Now the Grammar of Graphics is a description of how

kind of graphics can be broken down into abstract concepts.

You can, so think of the grammar of a language like English.

You have things like verbs and nouns and adjectives.

And so

the question is, you know, what are the

verbs, nouns, and adjectives of a data graphic?

And the Grammar of Graphics kind of describes kind of those basic elements

so that you can put them together to make new types of graphics.

Just like you could take a verb and a noun and an

adjective and make a new sentence that maybe no one's ever heard before.

You could take the grammar of graphics and put together various aspects

of plots and make a graphic that no one's ever seen before and

so that's the basic idea.

It's a very powerful concept to kind of organize all kinds of data graphics.

And until recently there was no specific implementation for it

in R, but Hadley Wickham who when he was a graduate

student at Iowa State implement the Grammar of Graphics as an

R package called ggplot and its current implementation is called ggplot2.

So one could think of this as almost

a third graphic system in R.

Even though it's based, is built upon the

grid graphic system which is, which comes with R.

It's kind of a third mode of, of plotting that has become very popular.

So if you think of the first mode as like

the base plots using functions like plot, and hist, and

boxplot, and then the second mode as the lattice plots

so using XY plot and these kinds of trellis type functions.

And then the third mode is ggplot.

So you get the package from CRAN. You can, you can use install.packages.

It installs on all almost all sys...I imagine on all systems.

You can go to the ggplot website which is ggplot2.org.

And so the nice thing about ggplot is that, is that, is that it is based

on this grammar of graphics, and so, it,

in a sense, there's a theory of the graphics.

So you can take this theory and kind of

reassemble the different pieces to make new types of plots.

And as Hadley Whitcomb says in his book, you know, the basic idea

is that you want to shorten the distance from the mind to the page.

So if you have some data that you're looking at, And you want,

and you thin of a way that you want to visualize that data.

You want to be able to rapidly take those

ideas and turn them into a picture on your screen.

So, from the GG plot two book,

this sentence kind of summarizes the basic idea.

But the idea is that, the grammar tells us

the statistical graphic is mapping from data to aesthetic attributes.

So color, shape and size.

- of geometric objects, so, points, lines, and bars, and the plot may also

contain statistical transformations of the data and

is drawn on a specific coordinate system.

So, we have things are that, we have a

mapping from data to aesthetics, geometric objects, we have statistics.

Now we have a coordinate system.

So in this lecture I just want to talk about

the qplot function which is kind of the most basic

function and it's probably the best place to start for

someone who is transitioning from say the base plotting system.

So in the base plotting system you know the work horse function

is the plot function and so qplot which you can think of as

standing for quick plot Is kind the work horse function for for

GD plot and its analogous to the plot function and the base system.

So one

key difference that you have to get used to when you're using

GD plot is that typically when you make a plot and you pass

data to the q plot function you want to tell it where the

data comes from and the data will always come from a data frame.

So a data frame is going to be.

So, your data have to be organized in a data frame.

And then when you plot variables those

variables are going to come from the data frame.

Now, you don't have to specify a data frame.

You can

if you don't specify a data frame the the cue plot function or

all the plotting functions will, will look for the data in your workspace.

But it's generally a good idea to specify the data frame.

That way when you read the code that generated

the plot You know exactly where the data came from.

So then so the data frame is

very important to organize before you start plotting.

Once you start plotting the plots are made up of aesthetics

and geoms and so aesthetics are things like the size, shape,

and color of things.

Points and the geoms are sort of

the objects that you're pointing, plotting I'm sorry.

So are you plotting points Are you plotting

lines, are you plotting bars, you know, whatnot.

One aspect that's important for the qplot function, and also is similarly

important when you're using lattice functions,

is the idea of using factor variables.

So factors are very important because they indicate subsets of your data.

So if you imagine you have a data frame or you have a y variable and a x

variable and then a factor variable the factor will

indicate subsets of your data in the data frame.

So for example you might have factor that indicates the gender.

So you have a bunch of

males and a bunch of females.

So those are subsets of your data and you

might want to plot a certain relationship divided by

the various subsets or you might want to color

certain points, depending on whether they're male of female.

And so the categories that are indicated by various

factor variables can be useful for annotating a plot.

And so, one aspect, so one thing that's

important about this feature is that, is that when

you have factor variables in a data set,

you want to make sure that they're properly labeled.

So it's usually not useful to label a factor variable

as one, two, and three, even if you have three categories.

One, two, and three is not particularly informative.

Usually you want to label them with the more informative labels

so that you know what those factor variables are trying to encode.

Now the qplot function is a fairly straight forward function

to use. I think it's very easy to pick up.

It hides a lot of the details of, of what

ggplot is doing underneath which is fine for many cases.

But the ggplot function is really kind of the core function of the system.

It's very flexible and you can use it in combination

with a lot of things that g, that qplot can't do.

Explore our Catalog

Join for free and get personalized recommendations, updates and offers.

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2019 Coursera Inc. All rights reserved.

Download on the App Store

Get it on Google Play