Yeah, I agree that SpaceX has access to much more data, but from reading the paper it appears that you can get quite far if your goal is to map atmospheric drag. I would suggest skimming through it if you're interested in the practical challenges.
> how long a prograde burn lasts, how much fuel it consumes, how much calculated delta-v each burn imparts to a satellite
Yes, to the first two, but surprisingly not to the third. Since the maneuvers are short and much higher thrust than the average atmospheric drag, you can accurately pinpoint where those maneuvers are occurring. The energy gain can only come from the thrusters, and you can also use your atmospheric model to subtract out the (very small, comparatively) energy lost to drag. So you can actually calculate the delta-V of each maneuver quite accurately even if you don't know its exact duration or propellant consumption.
> One of the most interesting satellites in recent memory is one that was specifically designed to be long and somewhat streamlined to prolong its lifespan at very low orbital altitude
GOCE was an astounding mission! They were trying to accomplish something much harder though, mapping tiny distortions of the Earth's gravitational field, which required an ion thruster to precisely cancel out the fluctuations due to atmospheric drag and maintain an inertial trajectory. You need several orders of magnitude less precision to map the atmospheric drag itself.
I think the coolest aspect of this work is that without any involvement from SpaceX, and taking advantage of the large numbers of Starlink satellites, the researchers were able to use some nice modelling and statistics to get some very high-quality data.