Con ten ts
01 Introduction to programming in R
02 Lecture (not in summary):
03 Hypothesis testing
04 Biostatistics – continuous outcomes
05 Binary outcomes
06 Survival analysis
07 Confounding and modifying
08 Prediction models
09 Repeated measures
10 Multiple testing
11 Overall summary: tests
12 Practical guide; How to perform tests in R
Data science · study summary
, lecture 01
Intr oduction to pr ogr am m ing in R
Data Science & Bioinformatics
Overview of R
Definition: R is described as a language and environment for statistical computing and graphics, available as a
free, open-source package compatible with all major operating systems.
Capabilities: It includes basic statistical procedures and allows for extensions through additional packages.
Statistical Analysis Stages
The presentation outlines five stages in statistical analysis:
Importing Data: Bringing data into R for analysis.
Data Inspection: Identifying errors and cleaning data, including recoding and summarizing.
Analysis: Estimating parameters, assessing uncertainty (confidence intervals, p-values), and determining
predictive values.
Model Validation: Checking assumptions of the model used.
Reporting Results: Summarizing findings through tables and graphics.
Basic Syntax and Data Structures
Data Structures: Key structures include vectors, matrices, modes, and data frames.
Vectors: Defined using c(), allowing for indexing and operations on elements.
Matrices: Two-dimensional structures that can be indexed similarly to vectors.
Data Frames: Tables that allow mixed data types, where each column can represent different modes.
Help pages always have the same lay-out:
Description -> usage -> arguments -> value -> details -> example.
Columns run vertically, rows run horizontally. In R, you specify [row,column] (row before column).
Giving names to variables
Be aware when giving variables names, no space or special characters! Also, numbers are not allowed as the first
characters. _ and . can be used to separate names. Names are case sensitive.
R has several modes:
Numeric -> 1,2,3,4, etc
Logical -> Boolean values (TRUE, FALSE)
== : test for equality and != : test for inequality.
Booleans are converted to numeric values. TRUE equals 1, FALSE equals 0. By applying sum, you can count
the numbers 1, which is the same as counting all the TRUEs.
Character -> between “.
Data science · study summary
, lecture 02
Lectur e (not in sum m ar y):
Data Science & Bioinformatics
Ctrl L = clear console.
Ctrl R = run script.
Would like to store your value in a certain variable, so you can use it e.g. as the argument of a function. In R, do
this with <-.
Functions are really important. The inputs are specified via arguments of the function between ():
Name_of_function(argument_1, argument_2).
Functions are in a general part of a package. Library() shows the packages installed on your computer.
Help-packages:
library(foreign) to make the package visible for your laptop. Read.dta was already installed, but r couldn’t see it.
Now, it can.
Trimmed mean
If e.g. trim=0.2, the 20% lowest and 20% highest values are left out and then the mean is calculated. Can leave
out the extreme values at the low and high end.
Help page:
Always same lay-out. Description usage arguments value details example.
Often you work with a series of numbers. How to make a single “factor” (e.g. patient name, patient number, drug,
etc). use the function c to make a vector:
> x <- c(10, 9, 8, 7, 6, 5, 4, 3, 2, 1)
Always different ways to do it:
x <- seq(from = 10, to = 1, by = -1) OR
x <- seq(10, 1) OR
x <- 10:1
if you don’t know the function very well, including the names in the function (like the first different way) can
already help you to understand. Only if you really understand it, start leaving more out like x <- 10:1.
# allows you to make comments in your script that wont be run.
Quite often you need to make a selection from a certain variable. E.g. separate analysis on patients over and
under 18. To extract an element from the data, you use []. First indicate the vector, and then the element you
want to extract. E.g. x[10]. To extract multiple elements, you first need to make a vector of the elements you want
to extract. indx <- c(5,10). Then do x[indx] to extract the elements 5 and 10 from the vector x. can also
do it like this:
c(-5, -10)
x[c(-5, -10)]
A matrix has 2 dimensions (rows and columns). Usage:
matrix(data = NA, nrow = 1, ncol = 1, byrow = FALSE, dimnames = NULL)
can either put the numbers in it by column or by row. By column means we start in the left upper corner and then
go down. In the example in the code this is the case. If byrow= FALSE, the numbers are put in the matrix by
column (default). If byrow = TRUE, the numbers are put in the matrix by row (left to right, then down).
[row,column]
Giving names to variables
Data science · study summary
, Often better for your code to use names that actually mean something. Be aware, no space or special characters!
Also, numbers are not allowed as the first characters. _ and . can be used to separate names. Names are case
sensitive. A and a are different variables. Also goes for Data and data.
Summary
Will give you a short summary of the variable. Useful to get a feel of the data you are working with, especially if
you import data from somewhere else.
R has several modes:
Numeric -> 1,2,3,4, etc
Logical -> Boolean values (TRUE, FALSE)
Very important. Can use TRUE/FALSE to select data. Can make a vector from e.g. numbers between 0 to 10
randomly. Then select all the ones greater than 5. Will get a list with true or false. Then, can select on the
trues to get all the values above 5.
== : test for equality and != : test for inequality. %in% c(3,8) to make a list of trues and falses in the
vector to see which ones are equal to 3 and 8.
Booleans are converted to numeric values. TRUE equals 1, FALSE equals 0. By applying sum, you can count
the numbers 1, which is the same as counting all the TRUEs.
Can also select on 2 (or more) valuables instead of just 1 (e.g. age and gender).
&: AND - all must be true
|: OR - at least one must be true
!: NOT - negation
Character between “.
Data frames
Columns and rows. Columns now have names. Better to use this in the code to make it more clear. Use names
between “ instead of numbers.
titanic[c(2,3),] # Or by name
titanic[,c("name","age")]
can also extract data for e.g. the variable age by using $:
titanic$age
dim(titanic) will give you the dimensions (number of rows and columns) of the data frame.
head(titanic) will give you the first few rows of the data frame.
tail(titanic) will give you the last few rows of the data frame.
To install a package that you don’t have yet, use the function install.packages("package")
install.packages("dplyr") and arrange function can arrange data.
E.g. titanic.SortedAge <- arrange(titanic3, age) that will sort on age from high to low.
To save a variable/data set, do this:
save(name dataset, file=”name.RData”)
E.g. save(titatnic3, file=”Titanic.RData’)
getwd() to get the working directory (where R is looking for/storing the data.
rm(name variable) to get rid of 1 of the variables in your working space.
Can use setwd() to set the working directory right, go to session in the menu and set working directory, or replace
the file to the working directory you’re in.
is.na(name) to see where the missing values are.
mean(name, na.rm = TRUE) to exclude the missing values in the table from the calculations.
Subset makes the selection a lot shorter.
Data science · study summary
01 Introduction to programming in R
02 Lecture (not in summary):
03 Hypothesis testing
04 Biostatistics – continuous outcomes
05 Binary outcomes
06 Survival analysis
07 Confounding and modifying
08 Prediction models
09 Repeated measures
10 Multiple testing
11 Overall summary: tests
12 Practical guide; How to perform tests in R
Data science · study summary
, lecture 01
Intr oduction to pr ogr am m ing in R
Data Science & Bioinformatics
Overview of R
Definition: R is described as a language and environment for statistical computing and graphics, available as a
free, open-source package compatible with all major operating systems.
Capabilities: It includes basic statistical procedures and allows for extensions through additional packages.
Statistical Analysis Stages
The presentation outlines five stages in statistical analysis:
Importing Data: Bringing data into R for analysis.
Data Inspection: Identifying errors and cleaning data, including recoding and summarizing.
Analysis: Estimating parameters, assessing uncertainty (confidence intervals, p-values), and determining
predictive values.
Model Validation: Checking assumptions of the model used.
Reporting Results: Summarizing findings through tables and graphics.
Basic Syntax and Data Structures
Data Structures: Key structures include vectors, matrices, modes, and data frames.
Vectors: Defined using c(), allowing for indexing and operations on elements.
Matrices: Two-dimensional structures that can be indexed similarly to vectors.
Data Frames: Tables that allow mixed data types, where each column can represent different modes.
Help pages always have the same lay-out:
Description -> usage -> arguments -> value -> details -> example.
Columns run vertically, rows run horizontally. In R, you specify [row,column] (row before column).
Giving names to variables
Be aware when giving variables names, no space or special characters! Also, numbers are not allowed as the first
characters. _ and . can be used to separate names. Names are case sensitive.
R has several modes:
Numeric -> 1,2,3,4, etc
Logical -> Boolean values (TRUE, FALSE)
== : test for equality and != : test for inequality.
Booleans are converted to numeric values. TRUE equals 1, FALSE equals 0. By applying sum, you can count
the numbers 1, which is the same as counting all the TRUEs.
Character -> between “.
Data science · study summary
, lecture 02
Lectur e (not in sum m ar y):
Data Science & Bioinformatics
Ctrl L = clear console.
Ctrl R = run script.
Would like to store your value in a certain variable, so you can use it e.g. as the argument of a function. In R, do
this with <-.
Functions are really important. The inputs are specified via arguments of the function between ():
Name_of_function(argument_1, argument_2).
Functions are in a general part of a package. Library() shows the packages installed on your computer.
Help-packages:
library(foreign) to make the package visible for your laptop. Read.dta was already installed, but r couldn’t see it.
Now, it can.
Trimmed mean
If e.g. trim=0.2, the 20% lowest and 20% highest values are left out and then the mean is calculated. Can leave
out the extreme values at the low and high end.
Help page:
Always same lay-out. Description usage arguments value details example.
Often you work with a series of numbers. How to make a single “factor” (e.g. patient name, patient number, drug,
etc). use the function c to make a vector:
> x <- c(10, 9, 8, 7, 6, 5, 4, 3, 2, 1)
Always different ways to do it:
x <- seq(from = 10, to = 1, by = -1) OR
x <- seq(10, 1) OR
x <- 10:1
if you don’t know the function very well, including the names in the function (like the first different way) can
already help you to understand. Only if you really understand it, start leaving more out like x <- 10:1.
# allows you to make comments in your script that wont be run.
Quite often you need to make a selection from a certain variable. E.g. separate analysis on patients over and
under 18. To extract an element from the data, you use []. First indicate the vector, and then the element you
want to extract. E.g. x[10]. To extract multiple elements, you first need to make a vector of the elements you want
to extract. indx <- c(5,10). Then do x[indx] to extract the elements 5 and 10 from the vector x. can also
do it like this:
c(-5, -10)
x[c(-5, -10)]
A matrix has 2 dimensions (rows and columns). Usage:
matrix(data = NA, nrow = 1, ncol = 1, byrow = FALSE, dimnames = NULL)
can either put the numbers in it by column or by row. By column means we start in the left upper corner and then
go down. In the example in the code this is the case. If byrow= FALSE, the numbers are put in the matrix by
column (default). If byrow = TRUE, the numbers are put in the matrix by row (left to right, then down).
[row,column]
Giving names to variables
Data science · study summary
, Often better for your code to use names that actually mean something. Be aware, no space or special characters!
Also, numbers are not allowed as the first characters. _ and . can be used to separate names. Names are case
sensitive. A and a are different variables. Also goes for Data and data.
Summary
Will give you a short summary of the variable. Useful to get a feel of the data you are working with, especially if
you import data from somewhere else.
R has several modes:
Numeric -> 1,2,3,4, etc
Logical -> Boolean values (TRUE, FALSE)
Very important. Can use TRUE/FALSE to select data. Can make a vector from e.g. numbers between 0 to 10
randomly. Then select all the ones greater than 5. Will get a list with true or false. Then, can select on the
trues to get all the values above 5.
== : test for equality and != : test for inequality. %in% c(3,8) to make a list of trues and falses in the
vector to see which ones are equal to 3 and 8.
Booleans are converted to numeric values. TRUE equals 1, FALSE equals 0. By applying sum, you can count
the numbers 1, which is the same as counting all the TRUEs.
Can also select on 2 (or more) valuables instead of just 1 (e.g. age and gender).
&: AND - all must be true
|: OR - at least one must be true
!: NOT - negation
Character between “.
Data frames
Columns and rows. Columns now have names. Better to use this in the code to make it more clear. Use names
between “ instead of numbers.
titanic[c(2,3),] # Or by name
titanic[,c("name","age")]
can also extract data for e.g. the variable age by using $:
titanic$age
dim(titanic) will give you the dimensions (number of rows and columns) of the data frame.
head(titanic) will give you the first few rows of the data frame.
tail(titanic) will give you the last few rows of the data frame.
To install a package that you don’t have yet, use the function install.packages("package")
install.packages("dplyr") and arrange function can arrange data.
E.g. titanic.SortedAge <- arrange(titanic3, age) that will sort on age from high to low.
To save a variable/data set, do this:
save(name dataset, file=”name.RData”)
E.g. save(titatnic3, file=”Titanic.RData’)
getwd() to get the working directory (where R is looking for/storing the data.
rm(name variable) to get rid of 1 of the variables in your working space.
Can use setwd() to set the working directory right, go to session in the menu and set working directory, or replace
the file to the working directory you’re in.
is.na(name) to see where the missing values are.
mean(name, na.rm = TRUE) to exclude the missing values in the table from the calculations.
Subset makes the selection a lot shorter.
Data science · study summary
