Birds fold, twist, and spread their wings in flight, leading to dramatic force changes used for their highly manoeuvreble flight. Their finely-tuned shape changes are very challenging to record and study. As a result, little is known about how bird flight is controlled, limiting our bioinspired aircraft designs. Now with motion capture, we have recorded high volume data using captive Harris hawks performing natural flight behaviours. With large, accurate data sets of real morphing shapes we can now use data-driven unsupervised machine learning methods for analyses, including dynamic mode decomposition. Such methods differ from traditional kinematics and from black box AI, leading to both explainable and intuitive results. From these large scale results we can resolve the morphing shape space, model flapping flight dynamics, and can now study the individual techniques and variation of coordinated flight controls.