Description
Matlab without plots is boring. Matlab with bad plots is worse. Matlab with good plots is… sublime. Since Fred’s been complaining about people’s
general plotting skills, I decided to spend some time on it explicitly this year.
Plotting well is as much an art as a science, so I’ll leave you with a few guidelines
and loose requirements. Get creative!
Here’s what each plot should have (even if not stated):
• a title (for subplots too!)
• axis labels (if none given, make some up label as x/y/z, or, if you’re really
brave, leave blank)
• a legend, if multiple datasets are overlaid on the same plot
• sane axis ticks and tick labels (set yourself if matlab’s autogeneration is
not sufficient)
• a sane color scheme (usually matlab does decently well here)
• a sane aspect ratio (default is usually fine but occasionally terrible)
These guidelines apply to subsequent assignments as well. Make me pretty plots.
😉
1 The Assignment
The following instructions are negotiable; if you don’t follow them exactly but
make a beautiful and readable plot anyway, I’m willing to be persuaded. I don’t
claim to know what the Fontaine wishes.
Two Dimensions Create the following plots. Don’t forget to call figure before
each to ensure you don’t overwrite previous plots.
1. Sample a sine wave at 50 points evenly spaced between zero and 2π.
Use plot to plot this wave in green.
2. Sample a cosine wave at the same 50 points and create a plot with
both waves superimposed (either use hold or pass matlabplot a
matrix – it can be done in one call!). Use legend to say which wave
is which, xlim & ylim to set (reasonable) axis limits, and xticks
& xticklabels to place labeled ticks at 0, π/2, π, 3π/2, & 2π. Give
the plot a terrible title (bonus points if you make me laugh).
3. Re-plot the sine and cosine waves in two different subplots using
stem (to indicate they’re discrete-time signals). Randomly change
the amplitude of one of them (see rand or randn to generate random numbers). Compute the amplitude ratio of the two waves in
decibels and use sprintf to embed that information in the title of
one of the subplots.
Set up the axes, ticks, titles, legends, colors, etc. of these plots as you
see fit. Turn the grid on for at least one of them.
4. Find a color image on the internet and download it. Use imread to
read it into matlab, and show it with imshow. Cool.
5. Examine the data that imread returned. Either use whos or the
workspace inspector – just get an idea of what type (integer/floating
point) and what size/shape the data is. Using this info, create another image (either mathematically or by modifying the returned
data) and plot that using imshow. Check out the docs for these
functions for more info on how they work.
Three Dimensions Create the following plots. Basically the same instructions,
but now we have an extra dimension to deal with!
1. Plot the surface z = exp(−x
2 −y
2
) over the square 0 ⩽ x ⩽ 1, 0 ⩽
y ⩽ 1. Create this plot three times, in three different subplots, using
plot3 (a strange choice here), scatter3, surf, and mesh. Adapt
the code in the lecture notes for this – the surface plotted there is
very similar.
2. Plot the parametric conical helix x = t cos 27t, y = tsin 27t, z = t,
for t from 0 to 1. Use pbaspect and/or daspect to set the aspect
ratio of the plot to 1:1:1, so that the cone looks nice. Set grid on
to give some background. Use view to set the camera position to a
nice view of the cone.
3. Read the code in the lecture notes for creating the matlab logo
and run it! This won’t be collected (though you can include it if
you wish) – it’s just for your amusement. It also uses a different
plotting approach (object-oriented rather than imperative), and as
such provides another potentially useful perspective.
A Creating a General Plot
This is mostly a catalog of how I do it, and as such should be taken with a grain
of salt.
I begin plots by calling figure. This creates a new plotting window, ensuring I don’t overwrite a plot I already made. I then set up subplots, if necessary
– I treat each call to subplot like the original figure call, so that nothing
gets overwritten in the subplots either. If I need to plot overlapping data, I call
hold on, which prevents subsequent calls to plotting functions from overwriting the plot window completely.
After that comes data – this roughly establishes what sort of axis scaling I’ll
need and lets me see where the plot needs work. plot, plot3, imshow, hist,
etc. – you get the idea. Once the data’s in place, I’ll title the plot, label the
axes, and add whatever legend I need, so I remember what the data on the plot
represents.
Finally, I’ll fiddle with the axis ticks, axis tick labels, plot aspect ratio, camera
positioning, background grid, and so on – all those little details that bring a plot
from workable to truly readable. At the very end, if necessary, I’ll export the plot
to an image using exportgraphics or something similar.
Not all of these steps are necessary for every plot, but if you want to produce
something polished, you’ll need many of them. Plus Fred will be happy. Also
your research partners and whoever’s reading your paper.
B Other Things To Check Out
• Mathworks has a useful intro to plotting – nothing fancy, but covers the
basics – at mathworks.com/help/matlab/learn_matlab/plots.html.
• The code in the lecture notes for the matlab logo was taken from Mathworks’ docs: mathworks.com/help/matlab/visualize/creating-the-matlablogo.html. I recommend looking at this page – it gives a beautiful livescript walkthrough of the process of creating the plot.
• The sprintf docs are at mathworks.com/help/matlab/ref/sprintf.html.
The format is very similar to C’s printf (as you probably guessed).