Hi guys.

New to running Stata for large groups. I'm working with IPUMs American Community Survey Data. I currently have person data for 38 cities in the U.S. and 13 years of data. In total, about 500,000 observations.

I want to create summary data for each city each year. There are many categorical variables "labforce" "race" "education" each with their own ranges. I want to create percentages for each value, e.g. I want to find the percent in the labor force each year for each city, and the percent not in the labor force for that city that year, the sum of these two values is 1 (100%) and completely defines the number of people in the labor force and not in the labor force in a particular city in a particular year.

Also there are person weights in my data set. So far as I can surmise, the weights are to make the sample more representative of the population of the U.S. as a whole. I'm not sure I want to use the weights. I think the weights are based on decennial census data and forcing the annual ACS data to match the characteristics of that decennial census data. The problem is that at any point in time, the weights can be between 1-10 years old. Also, the weights are set to match the entire U.S., which isn't what I'm focusing on. Even so, I am interested in running it flat and with weights, just to see what the differences might be.

I think the correct way to do this is either using contract or using tabulate but I'm not having much luck getting what I want.