The easiest approach I can see is to mark the data rows as the region, and then use a regexp search and replace:
M-x replace-regexp
RET \(.*\),\(.*\),\(.*\),\(.*\)
RET :PROPERTIES:
C-q C-j :Name: \1
C-q C-j :Tel: \2
C-q C-j :Mobile: \3
C-q C-j :Fax: \4
C-q C-j :END:
RET
If you needed to do this for many variable CSV files with different headers and numbers of columns, then I would probably approach it with keyboard macros.
user310031's answer would make a good basis for that. The macro could narrow the buffer to each row, insert the header row above it, perform the csv-transpose
(which appears to require CSV mode) do the search+replace, add in the :PROPERTIES:
and :END:
lines, widen the buffer again, and leave point on the line before the next data row. Then just mark the remaining data rows as the region, and type C-x C-k r.