Technische Universitt Mnchen

Engineering Data Analysis with MATLAB

Anke Scherb

Engineering Risk Analysis Group

Technische Universitt Mnchen

Phone: 089 289 23013

E-mail: [email protected]
Technische Universitt Mnchen

Engineering Data Analysis with MATLAB

Elective course 6th semester BSc
2 ECTS

Lecture period: Tuesdays 25.04. 16.05.2017

Lecture time: 08:45-10:15 am
Tutorials: 10:15-11:45 am
Lecture room: N 1039
Course language: English

Register for the course in TUMonline maximum number of participants 20

Download MATLAB at https://matlab.rbg.tum.de/
Using your myTUM credentials
Guidance for installation is provided there
Please bring your laptops with MATLAB installed

2
Technische Universitt Mnchen

Lecture - contents

Introduction to MATLAB: Basic commands

Working with vectors/matrices
Basic programming tools

(Descriptive) Statistics of data sets

Graphical representation of data sets
Statistics of pairs of data sets
Distributions and simulation of random variables

Exercises - contents

Practical exercises to be solved in MATLAB

3
Technische Universitt Mnchen

Examination: Project work

Groups of 2 students
Supervised by a tutor
Duration of project work: 6 weeks
Students must hand in a MATLAB code and a written report
Attendance of 3 out of 4 tutorials is required for participation

4
Technische Universitt Mnchen

Literature

Lecture Notes in Engineering Risk Analysis Lecture Notes in Engineering Risk Analysis
Part B Elementary Data Analysis Part Z Annex MATLAB
Prof. Dr. Daniel Straub Prof. Dr. Daniel Straub
TU Mnchen TU Mnchen

Probability Concepts in Engineering: Statistics, Probability and Reliability for

Emphasis on Applications to Civil and Civil and Environmental Engineers
Environmental Engineering N.T. Kottegoda and R. Rosso
A.H.-S. Ang and W.H. Tang McGraw-Hill, 1997
Wiley, 2006

http://www.mathworks.de/help/techdoc/

5
Technische Universitt Mnchen

Introduction to MATLAB

Why do we need programming at all?

MATLAB = MATrix LABoratory developed by MathWorks

Proprietary software

Main purpose: solve mathematical problems and present results with

graphical illustrations

Mainly used for numerical calculations/ simulations, and data acquisition/

analysis

Includes a number of useful toolboxes (e.g. statistics, optimization)

6
Technische Universitt Mnchen

Introduction to MATLAB

Advantages
Intuitive and easy to learn
Fast on matrix operations
Powerful syntax and great function library
Good plotting options
Easy debugging
Additional toolboxes with functions for specific purposes

Disadvantages
Interpreted language, no compiling
Not designed for word processing, internet applications, GUIs
Costs: commerically available

Free MATLAB alternative: octave

7
Technische Universitt Mnchen

MATLAB desktop working environment

Editor Editing
Workspace Collects
MATLAB programs
saved data

Current directory
Provides access to files

Command window Executes Command history Allows to

individual commands rerun previous commands

8
Technische Universitt Mnchen

MATLAB as calculator A very first example

=> Example.calculator

Variables Assigning values to variables

Name convention: Variable name = expression.
Start with letter Expression may be:
Letters, numbers, underscores Numerical value
Case sensitive Values combined with operators
Use meaningful names! Values of other variables
Returns from MATLAB functions
Returns from selfmade functions

Manual change in Workspace (go to

window Workspace and change value)

9
Technische Universitt Mnchen

MATLAB as calculator A first example

Task: Compute the moment of inertia of a rectangular beam section

bh3
Iy
12

1 clc h
2 clear all
3
4 % Input dimensions of beam in cm
5 b = 25; b
6 h = 40;
7
8 % Compute moment of inertia in cm^4
9 Iy = b*h^3;
10 Iy = Iy/12;

10
Technische Universitt Mnchen

MATLAB as calculator A first example

Task: Compute the moment of inertia of a rectangular beam section

bh3
Iy
12

Clears command window

1 clc h
2 clear all
3
4 % Input dimensions of beam in cm
5 b = 25; b
6 h = 40;
7
8 % Compute moment of inertia in cm^4
9 Iy = b*h^3;
10 Iy = Iy/12;

11
Technische Universitt Mnchen

MATLAB as calculator A first example

Task: Compute the moment of inertia of a rectangular beam section

bh3
Iy
12

Removes all variables

1 clc
from memory h
2 clear all
3
4 % Input dimensions of beam in cm
5 b = 25; b
6 h = 40;
7
8 % Compute moment of inertia in cm^4
9 Iy = b*h^3;
10 Iy = Iy/12;

12
Technische Universitt Mnchen

MATLAB as calculator A first example

Task: Compute the moment of inertia of a rectangular beam section

bh3
Iy
12

1 clc Comment line h

2 clear all
3
4 % Input dimensions of beam in cm
5 b = 25; b
6 h = 40;
7
8 % Compute moment of inertia in cm^4
9 Iy = b*h^3;
10 Iy = Iy/12;

13
Technische Universitt Mnchen

MATLAB as calculator A first example

Task: Compute the moment of inertia of a rectangular beam section

bh3
Iy
12

1 clc h
2 clear all
3
4 % Input dimensions of beamSuppresses
in cm screen
5 b = 25; printing b
6 h = 40;
7
8 % Compute moment of inertia in cm^4
9 Iy = b*h^3;
10 Iy = Iy/12;

14
Technische Universitt Mnchen

Basic commands

clc Clears the command window

clear all Removes all variables from memory

Removes variables var1 and var2 from
clear var1 var2
memory
save mydata Saves data to mydata.mat

load mydata Loads data from mydata.mat

whos Lists variables in memory and their sizes

; Suppresses screen printing

Continuous a line

% Designates a comment line

help fun Displays help topics about fun

15
Technische Universitt Mnchen

Vectors

Row vector Separate elements with spaces or commas

v = [3 1 7 -21];

Column vector Separate elements with semicolons ;

v = [3;1;7;-21];

16
Technische Universitt Mnchen

Vectors

Vectors with regularly spaced elements Colon sign :

v = [1:10] % from 1 to 10 with increments of 1 (default)

v =
1 2 3 4 5 6 7 8 9 10

v = [1:2:7] % from 1 to 7 with increments of 2

v =
1 3 5 7

v = [0.5:-0.1:-0.5] % from .5 to -.5; increments of -.1

v =
0.5000 0.4000 0.3000 0.2000 0.1000
0 -0.1000 -0.2000 -0.3000 -0.4000 -0.5000

17
Technische Universitt Mnchen

Vectors

Transpose operator Quote sign

>> v=[1:10] % Column vector

v =
1
2
3
4
5
6
7
8
9
10

18
Technische Universitt Mnchen

Vectors

Access vector elements

v(1)
u(3)
w = v(1:2:4); % vector elements 1 to 4 with increments of 2

Vector functions

length Computes number of elements

linspace Creates regularly spaced vector

Example: Vectors.m

19
Technische Universitt Mnchen

Matrices

Matrix Column of row vectors or a row of column vectors

A = [1 2 3;4 5 6;7 8 9];

B = [[1;2;3] [4;5;6] [7;8;9]];

Access row or column of a matrix

u = A(1,:); % access all column elements of row 1

v = A(:,2); % access all row elements of column 2

20
Technische Universitt Mnchen

Matrices

Matrix Add columns to Matrices

>> A = [1:3; 4:6; 7:9] % Or: A = reshape([1:9], 3, 3)

A =

1 2 3
4 5 6
7 8 9

>> A=[A,[-1;-1;-1]] % Add column to Matrix A

A =
1 2 3 -1
4 5 6 -1
7 8 9 -1

21
Technische Universitt Mnchen

Matrices

Matrix Add rows to Matrices/ Change Elements

>> A = [A;[-1,-1,-1,-1]] % add row
A =
1 2 3 -1
4 5 6 -1
7 8 9 -1
-1 -1 -1 -1

>> A(3,3) = 10 % change element

A =
1 2 3 -1
4 5 6 -1
7 8 10 -1
-1 -1 -1 -1

22
Technische Universitt Mnchen

Matrices

Matrix Change rows or columns

>> A(4,:) = [1, 1, 1, 1] % change last row Equivalent: A(4,:)=1
A =
1 2 3 -1
4 5 6 -1
7 8 10 -1
1 1 1 1

>> A(:,4) = [1; 1; 1; 1] % change last column

A =
1 2 3 1
4 5 6 1
7 8 10 1
1 1 1 1

23
Technische Universitt Mnchen

Matrices

Matrix Delete rows or columns

>> A(4,:) = [] % delete last row
A =
1 2 3 -1
4 5 6 -1
7 8 10 -1

>> A(:,4) = [] % delete last column

A =
1 2 3
4 5 6
7 8 10

24
Technische Universitt Mnchen

Matrices

Matrix operations All mathematical symbols perform matrix operations

A+B % Adds A and B

A-B % Subtracts B from A
A*B % Matrix multiplication
A/B % A*inv(B)

Element-wise operators Put a dot in front of a mathematical symbol to perform

operations on individual elements of the matrices. Matrices must be of same
size

A.*B
A./B
A.^B

25
Technische Universitt Mnchen

Matrices

Math functions of matrices Return matrices of same size with each entry
specified by performing the operation at the corresponding entry of the original
matrix
sin(A)
log(A)

Example: Matrices.m

26
Technische Universitt Mnchen

Matrices

Matrix functions - built-in functions

size Returns the size of each dimension of matrix

max Returns largest element of each column

min Returns smallest element of each column

sort Sorts each column

sum Sums each column

inv Computes the inverse of a matrix

Computes the transpose of a matrix

27
Technische Universitt Mnchen

Matrices

Special matrices: Identity Matrix

>> eye(3) % creates a 3x3 matrix

ans =
1 0 0
0 1 0
0 0 1

>> eye(2,3) % creates a 2x3 matrix

ans =
1 0 0
0 1 0

28
Technische Universitt Mnchen

Matrices

Special matrices: Zeros Matrix

>> zeros(3) % creates a 3x3 matrix

ans =
0 0 0
0 0 0
0 0 0

>> zeros(2,3) % creates a 2x3 matrix

ans =
0 0 0
0 0 0

29
Technische Universitt Mnchen

Matrices

Special matrices: Ones Matrix

>> ones(3) % creates a 3x3 matrix

ans =
1 1 1
1 1 1
1 1 1

>> ones(2,3) % creates a 2x3 matrix

ans =
1 1 1
1 1 1

30
Technische Universitt Mnchen

Saving the Workspace and loading data files

>> whos % whos displays in alphabetical order all variables in

the currently active workspace

>> save (myWS); % saves workspace to myWS.mat

>> load('myWS.mat'); % laods data of myWS.mat file;

31
Technische Universitt Mnchen

MATLAB script file versus command window

All of the pre-built commands that you use in MATLAB are script files or
functions (plot, mean, std, exp, cosd, )

MATLAB allows the user to create his/her own customized m-files for
specific applications or problems.

A script file is simply a collection of executable MATLAB commands. To

create a new script file, click on the New Script icon on the left side of the
Home Tab.

Save the file in an appropriate folder. When you pick a name for the file
you must follow the same rules that MATLAB has for naming variables.
Set the current directory in MATLAB to the same place where you saved
the script file.
To run the script file: Hit the Green Run Arrow in the Editor window

Example: CircleScript.m 32
Technische Universitt Mnchen

Elementary Data Analysis Descriptive statistics

All natural and human processes are subject to variability of some kind
=> hence there is most often uncertainty to some extent

Goal: Characterization and quantification of uncertainty to understand the

process

Descriptive statistics (data analysis)

Graphical representation of the data

Goal: Choice of appropriate probabilistic distributions and their parameters for

estalishment of probabilistic engineering models

33
Technische Universitt Mnchen

Elementary Data Analysis Descriptive statistics

Collection of experimental (measured) data/ samples

E.g. laboratory tests to determine water quality parameters, properties of

engineering materials, soil characteristics

Example: Tensile strength and Youngs modulus from tests on timber specimens
(provided by Lehrstuhl fr Holzbau und Baukonstruktion, TUM).

34
Technische Universitt Mnchen

Numerical summaries of data sets

Measures of central tendencies

Sample range Interval between the lowest and the largest value of the data set

range(data)

Sample mean arithmetic mean of the n sample values xi

1 n
x xi
n i 1
mean(data)

Note: The sample mean is sensitive to outliners (unexpectedly low or high values)

35
Technische Universitt Mnchen

Numerical summaries of data sets

Measures of central tendencies

Sample median The value for which there is an equal number of larger and
smaller samples

Sort data in ascending order

If the number of samples n is odd then the median is equal to the (n + 1)/2 th sample
If n is even then the median is obtained as the mean of the n/2 th and the (n/2 + 1) th
sample

median(data)

Note: The sample median is not affected by outliners

36
Technische Universitt Mnchen

Numerical summaries of data sets

Measures of dispersion (range, variance, standard deviation, interquantile range)

Sample variance Mean of the square of differences between the samples and
the sample mean
1 n
i
2
s
2
x x
n 1 i 1

var(data)

Note: The denominator n 1 leads to an unbiased estimator of the variance

37
Technische Universitt Mnchen

Numerical summaries of data sets

Measures of dispersion (range, variance, standard deviation, interquantile range)

Sample standard deviation Square root of the sample variance

1 n
i
2
s x x
n 1 i 1

std(data)

38
Technische Universitt Mnchen

Numerical summaries of data sets

Sample quantiles Values that separate the ordered samples into k groups of
equal number of samples. E.g. quartiles divide the samples into four groups of
n/4 samples.

Sample percentiles Values below which a certain percentage of all samples lie.
E.g. the 10-percentile x10 is the value below which 10% of samples lie

quantile(data,p)

prctile(data,p)

39