New to Matlab. I need help with group (?), double counting, and so on. Thank you!

New to Matlab, I need your help.

Suppose my data is this:

schoolid classid classsize enrollment

1001 1 10 22

1001 2 12 22

1005 1 14 29

1005 2 15 29

1018 1 8 22

1018 2 4 22

1018 3 10 22

.

.

.

And I want to find:

a. number of schools.

I found this one by

grps = findgroups(schoolid);

n = max(grps);

b. the average number of classes in a school.

I know how to solve by hand, but you know, my data is big,,, help me with codes please.

c. Provide the histogram of enrollment, and report its mean, median, max and min.

I know codes to find histogram, mean, and so on,

but you see the enrollemnt column is a double counting, I need to figure this out frist.

Please help me... Matlab geniuses... Thank you!

[Rik]

Have you looked at the splitapply function?

[JB]

I will look into it. Thanks!

Answer 1

[Image Analyst]

For a, and alternate way is

numberOfSchools = length(unique(schoolid))

A for loop way for b and c is

us = unique(schoolid)
numRows = length(schoolid)
for k = 1 : length(us)
    thisSchoolsRows = schoolid == us(k); % Only rows with this particular school
    numClasses(k) = length(thisSchoolsRows) % Assumes each classid's are unique and show up in only one row for a particular school.
end
aveClassesOverAllSchools = numClasses(k) / numberOfSchools

You can use the same approach to compute your other statistics. You can also use grpstats() or splitapply() like Rik mentioned.

[JB]

wow.... I hope I can write such code by my own someday. Thank you!