The way to measure downforce with suspension transducers is to drive the vehicle at a constant speed (or multiple set speeds) and measure the suspension deflection with the aero components. The surface needs to be flat and smooth though, with no accelerations occuring, so you can get a good steady reading for both sets of parameters. You can compare those suspension deflections to the same car at rest, or driven on the same road at the same speed with no aero, to get a good idea of the lift/downforce provided.
On a track, the suspension is constantly moving as the car pitches and rolls around every corner, and as each wheel reacts to every bump, dip, and curb it sees. Think about it - if the car goes over a crest and all 4 tires get airborne and fully extend, a suspension transducer-based measurement would think the aero package just made the car take flight!
Watching the video again, I would bet $ the downforce graph is being produced based on a formula off their speed data. Once you know downforce at a given speed, or the cL for your package, speed is all you really need to know, to know downforce. While the downforce vs. speed may not exactly match the curve I mentioned earlier ((speed^2)/22.5) due to small nuances in airflow throughout the vehicle's speed range, it is a sufficiently accurate approximation for something going into a Youtube video...
George, the devices you speak of, like suspension displacement transducers, or shock absorber load cells, are used by high-end race teams to help sort the suspension/chassis/aero packages of their race car. They might also be used by OEMs throughout vehicle development, though in that setting the goals are more compromised towards street civility.
High end race teams invest huge $$ in data acquisition gear, and will likely have a full-time team of "data guys" whose job is it to set up the sensors, collect and interpret the data, and turn that information into feedback the chassis team and driver can use to make improvements. This is in addition to all the data collected around engine and drivetrain operation, as well as the driver's throttle, steering, and brake inputs.
What I think is great is how some of this awesome technology has trickled down and become attainable at the amateur ranks. Today for a couple grand we can get compact data and video acquisitions systems (like the DL-1 or Chasecam systems) that in the Mark Donohue days, would have cost hundreds of thousands of dollars and filled multiple trucks/vans with bulky equipment.
