Is it possible to give an categorical date set values so it can be used for a plot.

I have a variable in my workspace: name=Wind_dir value=2560x1 categorical

These are wind directions (N, NNW, NW, NWW, ect) i can not figure out how to edit this. Is it possible to give like a value to N so i can use it for a plot or do i first need to de-categorical it or something like that. I dont have an idea where to begin.

[Cris LaPierre]

What do you envision your plot being?

[Dirk te Brake]

Something like a rose or compass.

Answer 1

[Dave B]

Note that you can plot categorical variables in a variety of ways. Passing your categoricals directly into your plotting function can be really useful!

c = categorical(["Apple" "Banana" "Orange"]);
y = [1 2 3];
nexttile;bar(c,y)

c = categorical(["Apple" "Banana" "Banana" "Banana" "Apple" "Orange" ...
    "Apple" "Orange" "Banana" "Banana" "Orange"]);
nexttile;histogram(c)
nexttile;pie(c)
nexttile;plot(c)

You can also convert categoricals to numeric...if you want. Calling double on a categorical will convert the first category to 1, the second category to 2, etc. Note that the order of categories typically depends on the second argument to categorical, or you can change it after the fact if you use reordercats. If you want to know the map between the categories and the numbers, you can use categories:

c = categorical(["Apple" "Orange" "Banana" "Banana"])

c = 1×4 categorical array
     Apple      Orange      Banana      Banana 

double(c)

ans = 1×4

1 3 2 2

categories(c)

You may also want to map some specific values for categoricals. You could define an equation to map them (in the case where they're ordered such that you can do some math from values like above:

c = categorical(["North" "North" "South" "South" "West" "West" ...
    "South" "East" "North" "North"], ["North" "East" "South" "West"])

c = 1×10 categorical array
     North      North      South      South      West      West      South      East      North      North 

(double(c)-1)*90

ans = 1×10

0 0 180 180 270 270 180 90 0 0

But in the more general case, you can look them up by defining a table (or two variables) that relates the categoricals to the numerics you want associated with them. Here's a 'fun' way to do that without a loop:

cardinal=categorical(["North";"East";"South";"West"]);
ang=[0;90;180;270];
t=table(cardinal,ang);
[r,~]=find(c==t.cardinal);
t.ang(r)

ans = 10×1

0 0 180 180 270 270 180 90 0 0

figure
polarhistogram(deg2rad(t.ang(r)),deg2rad(-45:90:315));
set(gca,'ThetaZeroLocation','top','ThetaDir','clockwise')

[Dirk te Brake]

Thank you very much! Very usefull information.

[Peter Perkins]

It's not widely known that, if c is a categorical, you can index into another array whose order is the same as c's categories. So IIUC, this step

[r,~]=find(c==t.cardinal);

in Dave's code would not be necessary. This

t.ang(c)

would do the trick.

[Dave B]

Unfortunately, it appears this trick doesn't work with tables, but you can use Peter's indexing magic with vectors. Though it's worth noting that if you do that you better make sure that they both specify the same order (which I skipped in my creation of cardinal above

c = categorical(["North" "North" "South" "South" "West" "West" ...
    "South" "East" "North" "North"], ["North" "East" "South" "West"]);
cardinal=categorical(["North";"East";"South";"West"], ["North";"East";"South";"West"]);
ang=[0;90;180;270];

ang(c)

ans = 10×1

0 0 180 180 270 270 180 90 0 0

% Unfortunately this isn't allowed:
t = table(cardinal, ang);
t.ang(c)

Error using tabular/dotParenReference (line 101)
A table row subscript must be a numeric array containing real positive integers, a logical array, a character vector, a string array, or a cell array of character vectors.