then wouldn't the decrease in rate of descent at the lower altitudes suggest an effort to recover from the descent?
Not IMO - the descent is "unremarkable" - an earlier poster referred to the FCTM and the descent rate is about half that of an Emergency Descent using Speedbrake.
The variations in V/S, including the reduction, I suspect are characteristic of IAS v M v AP mode, and the wind profile of the day. The corresponding fall in GSpd is a clue - Potential v Kinetic energy.
The data comes from FR24, or similar, using ADS-B. Of more interest would be adding the Mode S data downlinked to ATC e.g. IAS, Sel Alt etc. I would think that is available - just not in the public domain.