Base Plotting Demonstration

Loading...

From the course by Johns Hopkins University

Exploratory Data Analysis

3638 оценки

Johns Hopkins University

Exploratory Data Analysis

3638 оценки

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 1

This week covers the basics of analytic graphics and the base plotting system in R. We've also included some background material to help you install R if you haven't done so already.

Plotting Systems in R9:33

Base Plotting System (part 1)11:19

Base Plotting System (part 2)6:55

Base Plotting Demonstration16:56

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

So here I started up Bar, and I'm just going to do a little bit of

simple plotting just to kind show what the,

the basics are of the the plotting options.

So, I'm going to simulate a little, a little bit

of data here just so I can make a plot.

A hundred normal random variables here.

And I'm going to call Hist to make a histogram, of the plot.

So the first thing you'll notice when I call Hist, is that a plot window opens up.

So let me just move it over here.

And the

histogram of the data is shown.

And so you can see that even though I didn't specify any arguments in

the Hist besides the data themselves a number of things appear on the plot.

Without my specifying them.

So for example, there's a title here.

Then there's an x label here, so here the label is

just x because that's what the object called when I passed hist.

The label here is frequency which is the

default for histogram so it tells you the number of elements

within this range, so for example, between minus one and zero.

There's going to, there's a little over thirty elements of the vector in that

range, and you can see that the histogram is roughly like a normal distribution.

So let me just generate some more data here.

So we can make a little scatter plots I'm going to generate

some more data. And I'll plot xy here.

So now, the plotting window is already open, so when

I call plot it's not going to launch a new plotting window.

It's going to send the plot to the current plotting window, which is this one here.

So, I've speci-, I've called plot and you

can see the mixed scatter plot of the points.

The default plotting symbol here is an open circle.

You can see. And again

the label, I didn't specify besides x and y but

that a number of things have occured in the plot.

For example the label here is specified as x on the

x-axis, and the label on the y-axis is specified as y.

If I had changed the name of the objects, so let's

say I say [UNKNOWN] here, and I plot x and z.

Well the plot looks different because the data are different but you

can see that the label here turned, turned to Z because that's

the name of the object. So let me just go back to x and y.

And so a number of things on the plotting

region are important for example the margins here are.

There's four margins here, one for each side.

This is side one.

This is side two.

This is side three, and this is side four over here.

And you can see that the margin for the bottom is

is, is the largest.

So there's, there's five lines of margin text available.

On the side two, there's four lines of margin text available.

On the top, there's also four, so, side three.

And on the right side, side four, there's

the smallest amount of a margin text available.

So you can adjust that using the Mar function.

So for example, I can just say par, and then I say Mar equals, let's say

I want two on every side. So two, two, two, two.

Excuse me. And then I can plot again.

And you can see that now the plot the margins have gotten a

lot smaller and the plot kind of extends farther out into the window.

But the problem you can see is that on the x and the y axis, I lost my label.

And so even though I have the tick marks there I don't have my x and y labels.

So we probably

need to make them a little bit bigger than that.

So maybe I'll say four, four.

Like that.

So now if I plot again now you can see there's

just enough room for the x and the y label there.

I'm just going to demo a couple of the other options that

may be of interest to you as you're cri, constructing your plot.

So first is the plotting symbol I can say plot, x,y then PCH equals, say 20.

That gives me a solid circle here.

I want a slightly solid symbol here, I would say PCH equals 19.

Or I can, you know, specify PCH equals two, that gives me

triangles, three gives me little plus signs, four gives me x's, etcetera.

So you can see, so you can see that there's many

different plotting symbols to try, you might be wondering how I

know all the numbers for these plotting symbols, well, it just

comes after many years of use, I've memorize most of them all.

But, of course, if you haven't memorized them quite yet a

handy tool is the example file for points.

So if you just say points examples are points.

It'll go through a number of Demos so you

can see the capabilities that R can do with plotting.

But most important is their little plot

I'm sorry little chart of the symbols here.

And so you can see that's for example one is

the circle, two is the triangle, three is the plus etcetera.

Here at 20 was the solid circle small, 19 was the larger circle if you wanted the

solid square that's 15. Solid triangle is 17 etcetera.

So you can

specify what type of symbol you what and you just by, using the number here.

Another things, if you notice the symbols 21 through 25.

Those are actually symbols that are similar to say.

But to ones that are, have been previously shown, just for example, one thru six.

However the difference is in 21 through 25, those symbols have a

have have boundaries so you can see the boundary is red and they have a fill which

to this case is yellow.

So you can specify the two different colors, one for

the boundary of the outline and one for the color.

And so the boundary color specified using the col col argument.

And the, the background color, the fill color, is specified using the bg argument.

So you can specify two different colors like that if you

want.

Here's another chart of the plotting symbols, and

you can continue through the demo here for points if you want.

And I'll just finish it up here.

Just some of the special symbols that you can plot if you want.

This is in a different font and let me just click out of here.

So so that's one way to Change the symbol here.

I'm going to go back to my plot here. I'll say pch equals

let's say 20.

Excuse me.

I have to regenerate my data because it was overwritten by the example.

[NOISE] So here's my data again.

My little scatter plot.

Now, I can add a title to the plot by using the title.

Now I'll just say this is my scatter plot.

And that adds a title.

I could add some text if I wanted to, so for example,

I could use the text function to say add a little label here.

Let's see, I'll give it the coordinates I'll say minus 2 and then minus 2.

See that the label appears there.

I can add a legend if I wanted to. Say the legend.

And the legend, you can give it, kind of location specifications.

So, for example, top-left will put the legend in the top-left.

And then I'll say

I'll just give it a stupid label, excuse me let me

just pull plots in there so now I've got a legend there.

So there's all kinds of annotations that you can

add to the data as you kind of go along

so for example if I want to plot a line to the data I could fit a linear model.

A lot that using the LN function. Then the AD line function

will add the linear model fit on top of the here and the data aren't related

to the other, each other so that the linear model is the line is pretty flat.

If I wanted to adjust the thick, thickness of that

line, I can use the AD line and specify the LWD.

To be let's say three.

And you see that.

Now, a new line is plotted over that, which is much thicker.

So you probably wouldn't want to do this

from the get go.

If you wanted to remake this plot, you'd probably just

specify from the get go, LWD go up to three.

You don't want to necessarily plot two lines on top of each other.

But, I'm just showing this for demonstra- Demonstration sake.

I could have also changed the color.

So make the color, say blue. And that would make it blue.

And so there's lots of different types of options you can specify here.

When you're, as you're making the plot

So usually, you're going to want to create x labels and y

labels which are kind of, represent what the data are.

So you can plot, you can put a lot of these options in the plot function itself.

So I can say plot x,y and, and maybe x-lab is let's say, weight.

And then y-lab, or y-lab, will be, say, height.

The main will be scatter-plot.

PCH will be 20. So that gets.

Now, I have actual labels on the X-axis and Y-axis.

I can put my legend back if I want.

Maybe I'll put it in the top right this time.

And then I can add my little line here. So I can say, fit is,

[BLANK_AUDIO]

Or maybe I'll make the line red this time.

So, that's my plot, with the labels of the linear regression line.

And with the legend.

Now let's see what happens when we try to put more than one plot on the page.

So for example, let's say I have another variable, which I'll call z, and

maybe it's a, I don't know, maybe I'll make us some Poisson data here.

Oh, excuse me.

And let's say I want to plot z versus x and that

I also want to plot y versus x on the same canvas here.

So the first thing I can do is let's say I want to put,

let's say I want to put the plots right on top of each other.

So let's say par mf row equals so that it's going to happen, I

want to have two rows of plots and then one column of plots, right?

So that's what we want to see.

So now I'll apply the x and y on the top, and x and z on

the bottom. So x and y, so PCH equals 20.

So you can see that goes on the top, and on the bottom here,

I'll give x and z say it equals 19, and that goes on the bottom.

So now you can see that the margins are a

little bit large here, lar, probably larger than we would want.

and you see that the, that the margin that I have, that's specified here.

Is

that.

And so we want to adjust the size of the margin now

that we've kind of re- rearranged the configuration of the plots.

So we can maybe say something like

par, let's see, par equals

Now if I plotted again,

they get a little bit bigger.

So like 20 again.

So that's how, now I put two plots on the screen.

I could have done it the other way.

I could have said instead of them having top

and bottom, I could have had them right and left.

So by saying par name of row, equals a 1, 2.

Now I can plot like this

like that.

So you can see, I made the margins a little

bit too small, because I lost my y-axis label here.

So maybe I'll just say par.

I'll go back to four, four, two, two like this.

And make it positive

again.

So you can see that when you've rearranged the plotting layout, you might want

to rearrange the canvas itself to kind of remove some of the white space.

I won't do that for the moment, just so I can continue with the demo here.

But for example, you can put four plots on a page.

Like say, mar equals, sorry, mf row. Equals 2, 2.

That means two rows, two columns, so I can say plot x and y.

That'll go in the upper left. And you see, now, I can plot x and z.

You might wonder, well, where's the next plot going to go?

Well, because I specified mf row, the plots are going to go across the row.

So the next plot's going to be in the upper right.

And then the next plot's going to be in the lower left

so let's say I'll plot z and x instead of x and z.

And then the last plot's going to go in the lower right here.

So that, now I've got four plots on

a page by specifies, specifying the mf row option.

If I'd specified mf call the same thing would have happened but the order in which

the plots occurred Would have been different, so now I can say plot x, y, and

that appears in the same place, but the next

plot now is going to appear on the lower left, and

the next plot's going to be in the upper right,

and the last plot's going to be in the lower right.

So that's the difference between mf row and mf col.

The last option I'll talk about here is the points function, just as

a demonstration to how you can annotate a plot by adding things to it.

So, let me just reset the plot region so

that I'm only doing one plot at a time here.

Now suppose I generate some data and, and

suppose the data consists of say, men and women.

So there's two groups of people, here.

So I'm just going to generate some data here.

And maybe I'll give them a little relationship,

so it not, a little bit more interesting.

And let's say I've got another variable here which is the group and it's going to

be half and half so the two groups in the fifty level fifty iterations of each.

Let me just give them names.

We'll say labels.

So, I got males and females in this group of people here.

You can see it is a factor variable of two levels.

So, suppose I wanted to plot the data.

If I just plot the data X and Y you can't

tell who are the males and who are the females, right?

Because they're all the same color for example and So, suppose I want to plot the

data and plot the, make the males

one color and the females another Another color.

So how do I do that?

So the first thing you want to do, the basic idea is

you're going to set up the plot region, but

you're not going to plot any of the data.

And then you're going to add the data by gender, so you're going to

maybe add the females first, and then add the males, and the

idea is that each time you add the data points, they'll be

of a different color, or perhaps a different plotting symbol, or whatever.

So first, let's set up a plotting region, so I'm

going to say plot xy So I'm going to pass at the data.

But I'm going to say type = N.

So this means make the plot, but don't actually put the data in there.

So you can see that when I hit,

execute this function, everything happens just like before.

The labels are put in.

The tick marks are put in. The margins are specified.

Everything is there, except for the data.

So the only thing that's missing is the data.

And so what I'm going to do is, I'm going to add the

data, but I'm going to add them one group at a time.

So let's say I add the males first. So I can say points x, and then

I'm going to subset the vector so that the g, I

only take the points where g is equal to male.

Right, so that's a subset.

And then I'm going to say y, and then g is equal to male.

So this is only going to plot, plot the points, so where where the values of the

g variable are equal to male, let say I'm going to set, make the, that color green.

Okay.

So next I see the points on the, on the screen, are green.

Those are, those represent the male points only.

So I can do the same thing for the female, so

I can say points x and then g is equal to female.

And the y g is equal to female.

And let's say I'll make these, blue.

So now you can see that there are blue circles

for the females, and there's green circles for the males.

And so you can see the two groups, separately within the scatter plot.

And so, sub-setting based on a grouping variable is very common when making plots

and the points function can be used to kind of add points sequentially by group.

So that you can specify different types of properties for each group.

You can also, in addition to varying the

color I could have changed the plotting symbol.

So I could have said pch equal to let's say 19.

So this is a kind of solid circle here.

And that would have given me a solid blue circle

for the females and an open green circle for the males.

So that's one way to separate out.

Groups of data points on a single scatter plot.

Ознакомьтесь с нашим каталогом

Присоединяйтесь бесплатно и получайте персонализированные рекомендации, обновления и предложения.

Coursera делает лучшее в мире образование доступным каждому, предлагая онлайн-курсы от ведущих университетов и организаций.

© Coursera Inc., 2018 Все права защищены.

Загрузить из App Store

Загрузить в Google Play