Space weather events are associated with strong concentrations of magnetic field on the surface of the Sun. Currently, only low-fidelity space-weather forecasts are possible. Predicting the emergence time and size of the Sun’s active regions would be a significant step forward for space weather forecasting. The physics behind how these active regions emerge from the interior to the surface of the Sun is poorly understood. Only since the advent of a space-based monitoring campaign has it been possible to capture the emergence process of the magnetic field, Doppler velocity and intensity continuum of hundreds of active regions. By measuring the average motion of the polarities, surface velocities and pattern of convection, it is clear that convection plays an important role in the emergence of magnetic field on the Sun. Using machine learning and sophisticated simulations we aim to identify the convection cell dynamics associated with the emergence process, moving towards improved space weather forecasting.