The following code creates a adjacency matrix. For the example, students were asked to name up to five other students in their class that they considered to be friends. We start with a dataset that contains an id number for each case, as well as five variables that contain the id numbers of the friends each student identified. The code produces a square matrix with id numbers as row and column headings. The rows are which student was nominating friends, the columns are for nominees. A one indicates that the student for that row named the student for that column as a friend.
clear
input id friend1 friend2 friend3 friend4 friend5
44006 45611 55610 74158 55898 .
45611 44006 45623 45621 74158 55898
45621 71643 45611 . . .
45623 59642 71643 45611 73452 55610
55610 45623 45611 44006 55898 71643
55898 74158 55610 45621 . .
59642 55898 71643 45621 45611 74158
71643 55898 73452 59642 45611 44006
73452 71643 45623 . . .
74158 55898 45611 59642 45621 44006
end
* place variables to be in the resulting matrix in one matrix
* you will need to replace friend* with the appropriate variable names
mkmat friend*, mat(F)
* place the case id in another matrix
* you will need to replace id with the name of your id variable
mkmat id, mat(id)
* get number of rows in dataset
scalar define n = _N
* create fid, which is each case's row in the dataset, the data must remain
* (or be sorted) in the order defined by this fid
gen fid = _n
mata
/* get matrices from Stata into Mata*/
a=st_matrix("F")
id = st_matrix("id")
/* make an n by n matrix of zeros, where n is number of rows in dataset (cases)*/
b = J(st_numscalar("n"),st_numscalar("n"), 0)
/* loop through the rows and columns of the matrix b, setting the value in a cell
equal to 1 if the two cases are associated */
for(i = 1; i <= rows(a); i++) {
for(j = 1; j<=cols(a); j++) {
for(k=1; k<=rows(id); k++) {
if (st_data(k,"id")==a[i,j]) {
b[i, k]=1
}
}
}
}
/* add id as first column */
c = id,b
/* add id as first row with a dot in front */
c = (.\id)'\c
/* make this combined matrix accessible in Stata as "adj" */
st_matrix("adj", c)
/* exit mata */
end
mat list adj
/* matrix adj should look like this
adj[11,11]
c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11
r1 . 44006 45611 45621 45623 55610 55898 59642 71643 73452 74158
r2 44006 0 1 0 0 1 1 0 0 0 1
r3 45611 1 0 1 1 0 1 0 0 0 1
r4 45621 0 1 0 0 0 0 0 1 0 0
r5 45623 0 1 0 0 1 0 1 1 1 0
r6 55610 1 1 0 1 0 1 0 1 0 0
r7 55898 0 0 1 0 1 0 0 0 0 1
r8 59642 0 1 1 0 0 1 0 1 0 1
r9 71643 1 1 0 0 0 1 1 0 1 0
r10 73452 0 0 0 1 0 0 0 1 0 0
r11 74158 1 1 1 0 0 1 1 0 0 0
*/
preserve
set obs `=_N+1'
svmat adj, names(col)
* you may want to change temp.txt to another file name, you may also
* want to add a file path
outsheet c* using temp.txt, replace nonames
restore
The file temp.txt will contain the following matrix. Note that the rows are occupied by cases (as they were in the original dataset), while the columns correspond to the id numbers listed in the variables friend1-friend5. The matrix is not necessarily symmetric.
44006 45611 45621 45623 55610 55898 59642 71643 73452 74158 44006 0 1 0 0 1 1 0 0 0 1 45611 1 0 1 1 0 1 0 0 0 1 45621 0 1 0 0 0 0 0 1 0 0 45623 0 1 0 0 1 0 1 1 1 0 55610 1 1 0 1 0 1 0 1 0 0 55898 0 0 1 0 1 0 0 0 0 1 59642 0 1 1 0 0 1 0 1 0 1 71643 1 1 0 0 0 1 1 0 1 0 73452 0 0 0 1 0 0 0 1 0 0 74158 1 1 1 0 0 1 1 0 0 0