Bauer Vortrag Graphics

7/29/2019 Bauer Vortrag Graphics

1/49

IntroductionThe grammar of graphics

LayersDemo

ggplot2

Elegant graphics for Data Analysis

Johannes BauerSeminar (mit Bachelorarbeit)

20.1.2011

Johannes Bauer ggplot2 Elegant graphics for Data Analysis
http://find/


2/49


LayersDemo

Outline

1 Introduction

2 The grammar of graphics

The grammar of graphicsAesthetic attributesGeometric objectsFaceting

3 Layers

4 Demo

http://find/


3/49


LayersDemo

Introduction

ggplot2 is a plotting system for R

http://find/


4/49


LayersDemo

Introduction

ggplot2 is a plotting system for Rdeveloped by Hadley Wickham


I d i
http://find/


5/49


LayersDemo

Introduction


based on the grammar of graphics(Wilkenson, 2005)


I t d ti
http://find/


6/49


LayersDemo

Introduction


based on the grammar of graphics(Wilkenson, 2005)


Introd ction
http://find/


7/49


LayersDemo

Installation

install.packages(ggplot2)


Introduction The grammar of graphics
http://find/


8/49


LayersDemo


The grammar of graphics

The grammar tells us that a statistical graphic is a mapping fromdata to aesthetic attributes of geometric objects.


http://find/


9/49


LayersDemo


The grammar of graphics

aesthetic attributes aesthetics

colour

shapesize

geometric objects geoms

points

lines

bars


http://find/


10/49


LayersDemo


aesthetic attributes

example: qplot(carat, price, data=dsmall, colour=color)


http://find/


11/49


LayersDemo



example: qplot(carat, price, data=dsmall, colour=color)


http://find/http://goback/


12/49

The grammar of graphicsLayersDemo

g g pAesthetic attributesGeometric objectsFaceting


example: qplot(carat, price, data=dsmall, shape=cut)


http://find/


13/49


g gAesthetic attributesGeometric objectsFaceting


example: qplot(carat, price, data=dsmall, size=x*y*z)




14/49


Aesthetic attributesGeometric objectsFaceting

other aesthetic attributes

example: qplot(carat, price, data=diamonds)


IntroductionTh f hi

The grammar of graphicsA h i ib


15/49




example: qplot(carat, price, data=diamonds)


IntroductionTh f hi

The grammar of graphicsA th ti tt ib t


16/49




example: qplot(carat, price, data=diamonds, alpha=I(1/20))



The grammar of graphicsAesthetic attributes
http://find/


17/49




example: qplot(carat, price, data=diamonds, alpha=I(1/20))



http://find/


18/49



geometric objects

Geometric objects describe the type of object that is used todisplay the data.

some geoms

points (default)

smooth

bars

histogram

density

...





19/49



Adding a smoother to a scatterplot

example: qplot(carat, price, data=diamonds, geom=smooth)



http://find/


20/49




example: qplot(carat, price, data=diamonds, geom=smooth)



http://find/


21/49

g g pLayersDemo

Geometric objectsFaceting


example:qplot(carat, price, data=diamonds, geom=c(points, smooth))



http://find/


22/49

g g pLayersDemo



example:qplot(carat, price, data=diamonds, geom=c(points, smooth))



http://find/


23/49

LayersDemo



Methods



http://find/


24/49

LayersDemo



Methods

method = loess (default)uses a smooth local regression



L

The grammar of graphicsAesthetic attributesG i bj
http://find/


25/49

LayersDemo



Methods


method = lmfits a linear model



L

The grammar of graphicsAesthetic attributesG t i bj t
http://find/


26/49

LayersDemo



Methods



method = rlmworks like lm but uses a robust fitting algorithm so that

outliers dont affect the fit as much.



Layers

The grammar of graphicsAesthetic attributesGeometric objects
http://find/


27/49

LayersDemo



Methods



method = rlmworks like lm but uses a robust fitting algorithm so that

outliers dont affect the fit as much....



Layers

http://find/


28/49

LayersDemo


Faceting

With aesthetics we can compare different subgroups of ourdataframe on the same plot.



Layers

http://find/


29/49

LayersDemo


Faceting

With aesthetics we can compare different subgroups of ourdataframe on the same plot.

Faceting also splits the data in subgroups but displays it onmultiple plots.



Layers

http://find/


30/49

LayersDemo


Faceting

qplot(carat, data=diamonds, facets = color . ,geom = histogram, binwidth = 0.1, xlim=c(0,3))



Layers



31/49

yDemo

jFaceting

Faceting


Methodsfacets = color .



Layers

http://find/


32/49

yDemo

jFaceting

Faceting



geom = histogram has options like the binwidth



Layers

http://find/


33/49

Demo Faceting

Faceting




xlim=c(0,3) sets the limits of the x-axes



Layers

http://find/


34/49

Demo Faceting

Faceting


Methods

facets = color .


xlim=c(0,3) sets the limits of the x-axes

attributes xlim, ylim, main, xlab, ylab are equivalent to basicplot


IntroductionThe grammar of graphicsLayersD

The grammar of graphicsAesthetic attributesGeometric objectsF ti
http://find/


35/49

Demo Faceting

Faceting


IntroductionThe grammar of graphicsLayersDemo

http://find/


36/49

Demo Faceting

Differences from the basic plot function

qplot() is not generic.We cannot use points(), lines() or text() to add furthergraphic elements. We need additional layers.


http://find/


37/49

Demo

creating a plot

When we used qplot(), it did a lot of things for us:


http://find/


38/49

Demo

creating a plot

When we used qplot(), it did a lot of things for us:

it created a plot object

it added layers

displays the result

To create the plot ourself, we use ggplot().

ggplot() has two arguments

the dataframeaesthetic mapping


http://find/


39/49

Demo

creating a plot - example

Creates a plot but there is nothing to see.

p


40/49

Demo



p


41/49



p


42/49

creating a plot - complex example

p


43/49

creating a plot - complex example

p


44/49

creating a plot - simplified complex example

p


45/49


p


46/49


p


47/49

more examples in R


http://find/


48/49

References:

ggplot2 - Elegant Graphics for Data Analysis

Hadley Wickham, Springer 2009Online Referencehttp://had.co.nz/ggplot2


http://find/


49/49

Thank you.Questions?

http://find/

Documents

Bauer Vortrag Graphics