Cities represent a large set of cases which make them an excellent study object. Differences between parameters in cities like, say, composite indices of ambient air pollution (which can be somewhat objectively scored by looking at the state of knowledge regarding the extent of public health effects) can be caused by differences in public mores, legislation at the national level, geographical setting, and so on. You could call this a 'complex situation', but it does not make it any less simple or instructive to rank cities on their local air quality.

I am strongly in favour of using measurements that comprise empirical data as an input into decision making processes. You can't drive policy by deontic considerations alone, in my opinion. You will also have to aggregate some raw data for purposes of presentation, while minimising distortion and being transparent in how you do this.

That said, such real studies are possible, I think, with the technology and data we've gathered over recent years.  I think it's absolutely possible to quantify many of the attributes that make a city "good" and weight them according to a variable input, eg you have the inquirer submit to some survey questions ranking their preferred attributes then adjust the data weighting based upon that input and bam, you've got yourself a pretty good ranking list.  It can be used to draw some general conclusions but also account for differences between personal preferences.