Uplink (UL) electromagnetic field (EMF) exposure remains relatively underexplored, as most research has primarily focused on downlink (DL) exposure. Accurately assessing UL exposure is challenging due to adaptive power control mechanisms and the complexity of real-world propagation environments. This study investigates the correlation between user equipment (UE) transmission power and the distance to the nearest serving base station (BS). Field measurements were conducted across both indoor and outdoor environments in the Paris metropolitan area, covering four communication scenarios: traditional voice calls, WhatsApp voice calls, WhatsApp video calls, and FTP data transfers. By identifying the key influencing factors, this work supports the development of machine learning models for predicting UL EMF exposure levels more accurately.