Skip to Main Content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

University Library, University of Illinois at Urbana-Champaign

Engineering & Innovation

                   CIVL Home                                      Medical Students Home                                               Faculty Home

About Matlab

Good for:

  • Data analysis
  • Simple plotting
  • Parallel processing/bigger data

Why Matlab:

  • Excellent documentation
  • Extensible toolboxes offer wide capabilities
  • Used in academia for scientific study/analysis

Recommended setups:

 

You can install Octave on MacOS, Linux or Windows.  It can be run as a GUI (looks kind of similar to Matlab) or directly from the command line.

 

Best resource to learn quickly:

Matlab On-Ramp (create free account with illinois.edu email)

  • Matlab’s official on-ramp to working with Matlab.  Broken up by task/topic, these videos really help get started fast.  Mathworks/Matlab are known for having very good, useful documentation, making it easier to start building new projects with advanced functions/objectives.
  • Average video length:  ~10 min
  • Total duration: 2h15m

Loading Toolboxes

Toolboxes are loaded with the Matlab Installer, and have to be loaded before you launch Matlab. If you use the UIUC Virtual client (the one over Citrix), all the available toolboxes should be loaded for you. If you install your own version, you should be able to select at install which toolboxes to install. The default student version should install all toolboxes that comes with the installer.

Installing/Loading Packages with Octave

In Octave, additional functionality is achieved through the download and installation of Packages. Then, at each time you start Octave, you need to load the package much like a library in R or Python.

In [   ]:
# To install a package for Octave, download the package image to your computer.
# Then run:
pkg install pkgname-1.0.0.tar.gz;

# after the package installs, you use it as:
pkg load pkgname;

 

Using Multiple Matlab Scripts

How to use functions from other files

In [49]:
# to access a function from another script located in the same folder, 
# simply treat the name of the script like a function.
# Note: Only the first (outermost) function will be called from another script
# multiple functions can be defined in the other script, but they can't be accessed directly
# from another script. Look at the other_script.m file to understand this.

# You'll also get a warning the first time if the top function doesn't match the filename,
# but it will still run the first function no matter what it's called.

# In this example, other_script.m has two functions:
# other_function(a,b) and helper(n).
# other_function(a,b) calls function helper(n) to print out whatever 'a' was,
# then returns a + (b/2).  
# This is what happens when we call other_script like a function:
res = other_script(2,4)

# Note you can call it again too.
res2 = other_script(1,8)
 
you entered: 2
res =  4
you entered: 1
res2 =  5

Importing Data From CSV

How to import data into Matlab. Very useful for previewing and processing all kinds of information.

In [33]:
# For well organized, purely numerical data (NO TEXT), you have two choices.  
# For csv files containing numbers only, you can use csvread()
#csvdata = csvread('Path/to/file.csv', delimiter, rows, columns)


# If you have saved a variable/matrix to a .mat file, you can read it into Matlab with the load() function:
#matlab_data = load('path/to/mat/file.mat');

# if you are okay reading in without interpretation of data headers, just call import data on the file. 
data2 = importdata('data/ExampleDataClean2.csv');

data2(1:10)

# to read into a table/matrix with column labels, use 
#data = readtable('data/ExampleDataClean2.csv')
 
ans = scalar struct
ans =
{
  [1,1] = Timestamp,Val1,Val2,Val3
  [2,1] = 15:50:40.94
  [3,1] = 7741335 pig attached
  [4,1] = 15:50:41.08
  [5,1] = 7741335 pig attached
  [6,1] = 15:50:41.23
  [7,1] = 7741335 pig attached
  [8,1] = 15:50:41.38
  [9,1] = 7741335 pig attached
  [10,1] = 15:50:41.51
}

ans =

   -0.65625    6.69925
   -0.56250    6.77069
   -1.26562    7.15169
   -0.65625   10.19175
    1.06250   12.96194
   -1.29688   15.17650
   -0.35938   16.28775
    0.45312   17.90700
    0.26562   19.05000
   -0.25000   19.43100

ans =
{
  [1,1] = Timestamp,Val1,Val2,Val3
  [2,1] = 15:50:40.94,7741335 pig attached,-0.65625,6.69925
  [3,1] = 15:50:41.08,7741335 pig attached,-0.5625,6.7706875
  [4,1] = 15:50:41.23,7741335 pig attached,-1.265625,7.1516875
  [5,1] = 15:50:41.38,7741335 pig attached,-0.65625,10.19175
  [6,1] = 15:50:41.51,7741335 pig attached,1.0625,12.9619375
  [7,1] = 15:50:41.65,7741335 pig attached,-1.296875,15.1765
  [8,1] = 15:50:41.80,7741335 pig attached,-0.359375,16.28775
  [9,1] = 15:50:41.95,7741335 pig attached,0.453125,17.907
  [10,1] = 15:50:42.08,7741335 pig attached,0.265625,19.05
}

 
# For Octave, reading in the importdata() with headers allows Octave to determine numerical data from text.
# for this reason, it dumps the data in a struct which you can unpack as follows:
# if you have dirty data (not all numeric, unequal rows/columns), read in your data with importdata()
# the second parameter is the delimeter (CSV uses commas, TSV uses \t [tab], and still others use semicolons)
data = importdata('data/ExampleDataClean2.csv', ',', 1);
typeinfo(data)
all_fields = fieldnames(data)

data.textdata(1:10)
data.data(1:10)
 
ans = scalar struct
all_fields =
{
  [1,1] = data
  [2,1] = textdata
}

ans =
{
  [1,1] = Timestamp,Val1,Val2,Val3
  [2,1] = 15:50:40.94
  [3,1] = 7741335 pig attached
  [4,1] = 15:50:41.08
  [5,1] = 7741335 pig attached
  [6,1] = 15:50:41.23
  [7,1] = 7741335 pig attached
  [8,1] = 15:50:41.38
  [9,1] = 7741335 pig attached
  [10,1] = 15:50:41.51
}

ans =

 Columns 1 through 8:

  -0.65625  -0.56250  -1.26562  -0.65625   1.06250  -1.29688  -0.35938   0.45312

 Columns 9 and 10:

   0.26562  -0.25000

Importing Data from Excel

How to read in data from an excel data file

In [44]:
# In Octave, the Matlab 'table' datatype is not implemented. 
# This means mixed data (ie, text and numbers) takes a bit more effort to work with.
# An example is shown below that highlights how you 
pkg load io
pkg load dataframe

# in Matlab, you should use readtable() or readcell().  Although xlsread will still work as of 2020. 
A = xlsread('data/ExampleDataClean.xlsx');
typeinfo(A)
B = datestr(A(:,1), f=2);
C = [num2cell(B, 2), num2cell(A(:,2:end))];
D = dataframe(C);

A(1:5,:) # data read in.  If it can be converted to number it is. If only text, it shows up as NaN.
B(1:5,:) # first column converted to date strings.
C(1:5,:) # to blend text and numerical data, use a cell array.  each cell can hold different types of data
D(1:5,:) # there's also a dataframe in Octave, which is the closest thing currently to a Matlab table.
 
ans = matrix
ans =

 Columns 1 through 5:

   43957.000000            NaN       0.087072       0.466163       0.256585
   43958.000000            NaN       0.374148       0.596247       0.460567
   43960.000000            NaN       0.950790       0.786810       0.375649
   43962.000000            NaN       0.956982       0.137989       0.949380
   43964.000000            NaN       0.275037       0.759614       0.623671

 Columns 6 through 10:

       0.850972       0.024691       0.170840       0.306948       0.224802
       0.455865       0.473421       0.037419       0.164160       0.317844
       0.651548       0.224263       0.364410       0.277153       0.687860
       0.251682       0.422171       0.782476       0.938242       0.237547
       0.096792       0.265659       0.174680       0.850576       0.956483

 Columns 11 and 12:

       0.227992       0.423540
       0.357118       0.233253
       0.586596       0.062513
       0.456039       0.619374
       0.336973       0.135804

ans =

05/07/20
05/08/20
05/10/20
05/12/20
05/14/20

ans =
{
  [1,1] = 05/07/20
  [2,1] = 05/08/20
  [3,1] = 05/10/20
  [4,1] = 05/12/20
  [5,1] = 05/14/20
  [1,2] =  NaN
  [2,2] =  NaN
  [3,2] =  NaN
  [4,2] =  NaN
  [5,2] =  NaN
  [1,3] =  0.087072
  [2,3] =  0.37415
  [3,3] =  0.95079
  [4,3] =  0.95698
  [5,3] =  0.27504
  [1,4] =  0.46616
  [2,4] =  0.59625
  [3,4] =  0.78681
  [4,4] =  0.13799
  [5,4] =  0.75961
  [1,5] =  0.25658
  [2,5] =  0.46057
  [3,5] =  0.37565
  [4,5] =  0.94938
  [5,5] =  0.62367
  [1,6] =  0.85097
  [2,6] =  0.45587
  [3,6] =  0.65155
  [4,6] =  0.25168
  [5,6] =  0.096792
  [1,7] =  0.024691
  [2,7] =  0.47342
  [3,7] =  0.22426
  [4,7] =  0.42217
  [5,7] =  0.26566
  [1,8] =  0.17084
  [2,8] =  0.037419
  [3,8] =  0.36441
  [4,8] =  0.78248
  [5,8] =  0.17468
  [1,9] =  0.30695
  [2,9] =  0.16416
  [3,9] =  0.27715
  [4,9] =  0.93824
  [5,9] =  0.85058
  [1,10] =  0.22480
  [2,10] =  0.31784
  [3,10] =  0.68786
  [4,10] =  0.23755
  [5,10] =  0.95648
  [1,11] =  0.22799
  [2,11] =  0.35712
  [3,11] =  0.58660
  [4,11] =  0.45604
  [5,11] =  0.33697
  [1,12] =  0.42354
  [2,12] =  0.23325
  [3,12] =  0.062513
  [4,12] =  0.61937
  [5,12] =  0.13580
}

ans = dataframe with 5 rows and 12 columns                                                             
_1       C1     C2       C3      C4      C5       C6       C7       C8      C9     C10     C11      C12
Nr     char double   double  double  double   double   double   double  double  double  double   double
 1 05/07/20    NaN 0.087072 0.46616 0.25658 0.850972 0.024691 0.170840 0.30695 0.22480 0.22799 0.423540
 2 05/08/20    NaN 0.374148 0.59625 0.46057 0.455865 0.473421 0.037419 0.16416 0.31784 0.35712 0.233253
 3 05/10/20    NaN 0.950790 0.78681 0.37565 0.651548 0.224263 0.364410 0.27715 0.68786 0.58660 0.062513
 4 05/12/20    NaN 0.956982 0.13799 0.94938 0.251682 0.422171 0.782476 0.93824 0.23755 0.45604 0.619374
 5 05/14/20    NaN 0.275037 0.75961 0.62367 0.096792 0.265659 0.174680 0.85058 0.95648 0.33697 0.135804

Plotting Data

How to generate a basic data plot

In [63]:
# To plot data, use the plot() function.
# we'll demonstrate by plotting from a csv file.

pkg load io

A = xlsread('data/ExampleDataClean.xlsx');

# plot the third and fourth columns as time series
# first, generate the 'x' values. Our data are columns,
# so we need to match the dimensions, hence the transpose
x = transpose(linspace(1, 60, size(A, 1)));

# Now we plot the third and fourth columns.  
# The third column will be a blue line, and the second red stars.
plot(x, A(:,3), 'b-', x, A(:,4), 'r*')
 

Extra Tidbit

A fully functional code snippet that may be useful.

In [   ]:
# here is a function for reading in WFDB format files from physionet (raw EMG data, for example)

function [times, data] = readWFDB(filename)
    fid = fopen(filename);
    % get HEA info like offset, format, frequency, gain
    offset = 20;
    gain = 500;
    sample_frequency = 23437.5;
    precision = "int16";
    format = 'l'; % little endian
    data = fread(fid, Inf, precision, 0, format);
    data = data([offset:end],:);
    data(:,1) = data(:,1) ./ gain;
    len = size(data,1);
    times = linspace(0, len -1, len)';
    times = times./sample_frequency;
    fclose(fid);
end