Even cylindrical pressure vessels have spherical endcaps. Besides I have nested many cylinders *inside* the sphere, so the habitat is still essentially cylindrical in nature. All I've done is remove the cylindrical section separating the endcaps and bring the endcaps together to form a complete sphere. The reason I did this is to minimize the amount of uranium dioxide needed to maintain heat in case of power failure. Otherwise all water on board would freeze within 3 weeks. That is bad for any biota because freezing water expands and ruptures cell walls. This would also give the astronauts a safe haven while they get the main power supply repaired and running again. If repairs take months then they would have time to properly store food and live off that after plant life dies. Once power is back up they could replant.

The point is they would still be alive, not frozen. If you would like to try out your own starship habitat shapes for heat loss you can use the equation:

q = e . B . A . (Ti^4 - To^4)

where,
q = heat radiated in watts
e = emissivity of insulating material
B = 5.67E(-8) Watts/(m^2.deg.K)
A = surface area of radiating surface in square meters
Ti = inside temperature in degrees Kelvin
To = outside temperature in degrees Kelvin

Once you decide on a geometry you can use a spreadsheet like MS Excel, Lotus 123, or Corel Quatro Pro to work out the results fast.