Here is the plan. I need to delete every metro's least populated census tracts (in terms of population density). To what extent? As to preserve 98% of the metro's total population. I don't know how to program. Is there any existing tools I can use?
Thanks everybody!!
S.
Just sort the data by increasing density and remove records until up to 2% of the population has been eliminated. Here are the details if you're unfamiliar with this kind of analysis in ArcGIS:
Compute population density by dividing population by area. (If you don't have areas, compute them. It's best to use an equal-area projection, but for most cities the projection won't make much of a difference.)
Export three columns in a format readable by your favorite statistics or spreadsheet software: tract identifier, population, and density.
In your stats/spreadsheet software, sort the records by increasing density.
Compute the cumulative population: that is, starting with a previous population of zero, add the current population to the previous population.
Normalize the cumulative populations: divide them by the total population (found as the very last cumulative population).
Delete all records where this fraction is 1 - 98/100 = 0.02 or less.
Save the result in a format ArcGIS can read.
Back in ArcGIS, join the result to the original table based on tract id.
Use a definition query or a special symbol in the map to eliminate the records having nothing joined to them.
