Experimental epidemiology analyses with R and R commander Lars T. - - PowerPoint PPT Presentation

1

Experimental epidemiology analyses with R

and R commander

Lars T. Fadnes Centre for International Health University of Bergen

2

• Click to add an outline

3

How to install R commander?

• install.packages("Rcmdr", dependencies=TRUE)
• Download necessary web files put them in a folder:

http://statistics.fadnes.net/epi

– Fieldtrials12.RData – Exercises-for-the-exp-epi-2012-03.rtf

4

• Install R (if not done already)

– http://www.r-project.org/

• Install package Rcmdr (if not done already)

– Package Install package scroll down to Rcmdr click OK

• Install package Epi

– Package Install package scroll down to Rcmdr click OK

• Download RcmdrPlugin.Cih.Epi from

http://statistics.fadnes.net/epi/

• Install package(s) from local zip file (choose the package you just downloaded)
• Open program

– Load package Rcmdr – Tools -> load Rcmdr plug-in(s) RcmdrPlugin.Cih.Epi

• click ok and Yes
• Now you are ready

Installation of RcmdrPlugin.Cih.Epi

5

Aim for this session

• Introduce a brilliant tool
• Analyse a dataset with the tool
• Guide to further knowledge

6

What is R?

• R is a free software environment that

includes a set of base packages for graphics, math, and statistics.

• You can make use of specialized

packages contributed by R users or write your own new functions.

7

Why R?

• Very powerful
• Developing extremely quickly
• Working on different platforms (not only

Microsoft Windows…)

• Free of all costs

8

Why don’t all use R?

• Click to add an outline

9

What is R commander?

11

Why R commander?

• Powerful
• Free of all costs
• Working on different platforms (not only

Microsoft Windows…)

• Easy to learn and to use…

12

How to install R?

For Windows:

• http://cran.r-project.org/bin/windows/base/

– Easy

• If here at UiB, the IT department will fix it for you if you just ask them

to add it for you For Linux (Ubuntu etc):

• Very Easy
• Just search for R-base-core in Synaptic Package Manager and add it
• http://socserv.mcmaster.ca/jfox/Misc/Rcmdr/installation-notes.html

(also contains a good description for installation on Mac)

13

How to install R commander?

• Is already installed on UiB computers
• If not:

install.packages("Rcmdr", dependencies=TRUE)

14

Some things to note first

• R is case-sensitive

– help, Help, HELP and HELF are different… – Recommendation:

• Choose one style and stick to it
• If it’s something you don’t know?

– There are lot’s of good information on the web

• Particularly for R

15

How to start?

• Open R

– Load packages

Rcmdr

• r write
• library(Rcmdr)

16

Menu

• File Menu:

– items for loading and saving script files; – for saving output and the R workspace; – and for exiting

• Edit Menu:

– items (Cut, Copy, Paste, etc.) for editing the contents of the script and output windows.

• Data

– Submenus containing menu items for reading and manipulating data.

• Statistics

– Submenus containing menu items for a variety of basic statistical analyses.

17

Menu

• Graphs

– Menu items for creating simple statistical graphs.

• Models

– Menu items and submenus for obtaining numerical summaries, confidence intervals, hypothesis tests, diagnostics, and graphs for a statistical model, and for adding diagnostic quantities, such as residuals, to the data set.

• Distributions

– Probabilities, quantiles, and graphs of standard statistical distributions (to be used, for example, as a substitute for statistical tables) and samples from these distributions.

• Tools

– Menu items for loading R packages unrelated to the Rcmdr package (e.g., to access data saved in another package), and for setting some options.

• Help Menu

– items to obtain information about the R Commander (including this manual). As well, each R Commander dialog box has a Help button (see below).

18

• Script Window

– R commands generated by the R Commander – You can also type R commands directly into the script window or the R Console – The main purpose of the R Commander, however, is to avoid having to type commands.

• Output Window

– Printed output

• Messages Window

– Displays error messages, warnings, and notes

• Graphics Device window

– When you create graphs, these will appear in a separate window outside of the main R Commander window.

19

Available functions:

20

Click to add title

21

Save dataset under your documents folder

Files and documents are available at http://statistics.fadnes.net/epi

22

Let’s get started…

• Change directory (under File)

– Find the folder where you placed your data file

• Import data

– Give it the name: fieldtrials

• Save workspace as

– Give a name to your file – The file contains the dataset and

any models you might have generated

23

Data Types

• Vectors

– Quantitative difference (one vs. two apples) – including continuous (numerical) variables – Number variables coded as vectors as default

• Factors

– Qualitative difference (apples vs. pears) – Categorical – Text variables coded as factors as default

• Matrices, lists, arrays and data frames

http://www.statmethods.net/input/datatypes.html

24

Variables in dataset - define the datatypes

id id number gender male/female - (factor) treatmentarm Treatment (1=zinc, 0=placebo) - (factor) childage Age of the child in months - vector breastfed Is the child breast fed - (factor) lentils Does the child eat lentils? (0=no, 1= yes) - (factor) meat Does the child eat meat? (0=no, 1= yes) - (factor) duration Duration of diarrhea in days - vector diarsev Severe diarrhoea ≥10 stools per day - (factor) fever Did the child have fever at enrollment? - (factor) clusterzn2/4/8/16 cluster variables identifying living areas

• coded as vector, but needs to be transformed into a factor

25

How to save?

• Save R workspace as…

– This will save your data (in the R format)

• Save output as…

– This will save your output – Another strategy is to cut and paste what you want to save

• Always save the commands (syntax)

– essential if you want to re-run the analyses later – WordPad is a better option than Word etc (does not autocorrect - change to upper case etc)

26

How to write and a command?

• Simply write the command in the script

window, mark it and click ’Submit’ or press Ctrl+R

27

Nice to know:

• When writing comments in the syntax,

start with the following sign #

– R will then not consider the line as a command

• If you are uncertain about a function, use

google or help(name-of-function)

28

– Cluster has numbers and is as default coded as vector, but needs to be recoded into a factor (categorical variables for grouping etc)

29

Now we’re ready to answer some scientific questions…

30

Compute new variable

• Child age in years
• Child age now given in months
• of vaccination is often calculated by measuring

antibodies before and after vaccination

• childageyear = childage/ 12

31

Does the new variable look reasonable?

• View data set

32

Summarize variable

• Calculate mean, median and standard

deviation for

– childageyear

• for each intervention arm (treatmentarm)
• Numerical summaries

33

Doing calculations for subsets by generating new datasets

34

• Placebo:

treatmentarm == "placebo"

• Zinc:

treatmentarm == "zinc" Make the other subset by changing the syntax and run ('Submit') the syntax

zinc <- subset(fieldtrials3, subset=treatmentarm=="zinc") placebo <- subset(fieldtrials3, subset=treatmentarm=="placebo")

35

– You can now easily change between the datasets

36

Make a histogram of childageyear

• First for the 'zinc' dataset
• Then for the 'placebo' dataset
• Are they look similar?

Note: – The histograms will be printed in the R window (not inside R commander) – Right click on the graph and you can copy it as metafile to paste it into a document, print it or save it

37

• Does it look normally distributed?

placebo$childageyear frequency 0.5 1.0 1.5 2.0 2.5 10 20 30 40

38

Box plot

• Box plot for childageyear

– By treatmentarm – (first remember to select the complete fieldtrials dataset)

placebo zinc 0.5 1.0 1.5 2.0 2.5 treatmentarm childageyear

39

Is the baseline child age different in the zinc and placebo arms?

• This can be checked with a robust test not

assuming normal distribution?

• Check with a

– non-paramethric » two-sample wilcoxon test (log rank test) » Use the ’Exact’ test

40

Recoding variables

Diarrhoea duration can be recoded into an additional categorised variable (diarlong) – Data  manage variables

• value = ”factor”

– Factor can be either number or word

• value, value, value = ”factor”

– Listed with comma

• value:value = ”factor”

– From lowest to highest values

• else

all other values

• NA

missing

41

Click to add title

42

Are there differenses in syntax between R and R commander?

• Some few:

– Commands that extend over more than one line should have the second and subsequent lines indented by one or more spaces or tabs; all lines of a multiline command must be submitted simultaneously for execution. – Commands that include an assignment arrow (<-) will not generate printed

• utput, even if such output would normally appear had the command been

entered in the R Console [the command print(x <- 10), for example]. On the other hand, assignments made with the equals sign (=) produce printed output even when they normally would not (e.g., x = 10). – Commands that produce normally invisible output will occasionally cause output to be printed in the output window. This behaviour can be modified by editing the entries of the log-exceptions.txt file in the R Commander’s etc directory. – Blocks of commands enclosed by braces, i.e., {}, are not handled properly unless each command is terminated with a semicolon (;). This is poor R style, and implies that the script window is of limited use as a programming editor. For serious R programming, it would be preferable to use the script editor provided by the Windows version of R itself, or – even better – a programming editor.

43

True or false quiz

• R commander only works for Ms Windows?
• R is case sensitive (difference with small and large

letters)

• R is built by a few people with a secret source-code?
• R was the program that gave their shareholders most

profit last year?

• There are a lot of enthusiastic people working with R

providing help to their peers in the R forum?

44

R help forum:

• http://r.789695.n4.nabble.com/

45

Further reading:

• The R Commander A Basic-Statistics Graphical User Interface to R
• John Fox 2005.pdf

– http://www.jstatsoft.org/v14/i09/paper

• Getting started with the R Commander: a basic-statistics graphical

user interface to R

– http://socserv.mcmaster.ca/jfox/Getting-Started-with-the-Rcmdr.pdf

• Quick-R: magnificent guide

– http://www.statmethods.net/

• http://cran.r-project.org/manuals.html

46

• You have learnt some basic skills and can

now experiment with the program yourself

47

Questions and comments

• Click to add an outline