I will discuss the story behind the collection of wrist-worn accelerometer data in over 150,000 research participants across the UK and China, while also describing efforts to collect complementary open human activity recognition validation datasets to further enhance these resources. I will share the development of machine learning methods for sleep, sedentary behaviour, physical activity behaviours and steps, referring to open software tools and data resources of relevance to others in the field, and hopefully beyond to identify opportunities for collaboration with rare disease researchers.