Finding census block with highest pop. value within census tract using QGIS?

In QGIS, I'm trying to plot a centroid of only the census block with the highest population within each census tract. So there should only be one point per tract- that of the block with the greatest population. The two shapefiles that are relevant here are: 1. Census Tracts, which has the ID of each tract in the attribute table, and 2. A population density grid at the census block level, which has the pop. count per block, the tract ID, and the blocks ID in the attribute table. See screen shot. Thick lines are tracts, thin lines are blocks.

The end result is I want to do a nearest hub analysis based on the block with the highest population within any given tract.

I'm new to GIS, so you may have to spell some things out for me. I'm only about a month in.

• To help you further I need to understand exactly you mean when you write, "the highest population within each census tract". Which variable are you referring to? Is this just a raw count of population or does it include some other attribute of census data in aggregate form (income, education levels, language(s) spoken, etc.). When you further explain, "So there should only be one point per tract" makes me think its raw data and you want an XY point placed in the centroid of the tract. If so, this may be misleading as the centroid may not be exact location where the highest population is locate – logrady Dec 28 '14 at 3:45
• The variable I'm referring to is the block with the most population. No other variables are being considered here. The data was retrieved from TIGER. I think this would not be raw data, as Census staff probably did some calculations. Specifically, here's where I got the population data: census.gov/geo/maps-data/data/tiger-data.html > 2010 Census > Population & Housing Unit Counts - Blocks > Nevada. The XY point would be at the center of the block with the most population within any single tract. Therefore, any single tract should have just one XY point. – free_radicals Dec 28 '14 at 4:01
• I assume you are asking this in order to make progress on your other question gis.stackexchange.com/questions/127598/…. I have provided an extensive proposed answer to that question. Census blocks might be scattered in odd ways, and you can't assume that the centroid of the highest population block is a better approximation of the mean center of tract population than the tract centroid. Please look at my proposed answer to your other question and consider revising or deleting the current question. – Lee Hachadoorian Dec 29 '14 at 21:11