Taken from this paper on using remote sensing to measure urban agglomeration in two dimensions; they took ten years of data from two satellites, one measuring night time lights to measure sprawl, and another measuring microwave scattering to infer vertical build-up. They divided each city into a 21x21 grid, threw out the ones that contained just water, tallied the change in observed lights and microwave bouncin' (the build-out and build-up, respectively) for each square, and stuck a tiny line chart in each one to show the ten year change. Blue means not much changed, red means a lot was added. Left column NL is nighttime lights, and on the right PR stands for "power return," or the relative amount of energy bounced back to the satellite. A higher power return (augh augh augh) means more bigass buildings getting in the way and reflecting more microwaves.
Dense urban cores in Beijing and Shanghai didn't add much illuminance since they're already built up and nova-bright, but they sure as hell are growing vertically; just look at those ulcers in the right column. India's cities on the other hand aren't growing "up" much at all, but are creeping outwards in informal-as-hell ways. Neat! But I'd throw out the line graphs entirely and just code each square by color. You can't even make out the slopes of the lines, so why bother with anything but the hue?