A primary characteristic that these cities share is fixed land areas during the last 40 or more years (and for some of the cities, up to 100 or more years). So there population declines in older residential areas haven't been masked by annexation or growth in greenfield areas. A more useful chart is one replacing the population with the number of households. These data better reflect abandonment, and eliminate the population declines attributable to decreases in household size (which for Detroit is a factor that could account for at least 600,000 of the population decline). Even if the actual household count data aren't readily accessible, you can easily convert the city populations to households by using historical data for average household sizes in the US (which are available back to 1790).What you will find from these data are when the really declines attributable to abandonment started occurring (as well as the finding that Milwaukee's households have actually increased in number while the city's population was decreasing by >130,000 – which is a vastly different scenario than the assumed abandonment narrative supposedly associated with industrial decline).