I have an unbalanced panel data, and I'm trying to weight my sample to look like the population. suppose I have n city and each city has some observation during a period of time. Should I use pweight option in my regression? how?