Er clustering plays a major role in hindering sufficient optical gain in Er-doped Si materials. For porous Si, the long-standing failure to govern the clustering has been attributed to insufficient knowledge of the several, concomitant and complex processes occurring during the electrochemical Er-doping. We propose here an alternative road to solve the issue: instead of looking for an equilibrium between Er content and light emission using 1-2% Er, we propose to significantly increase the electrochemical doping level to reach the filling the porous silicon pores with luminescent Er-rich material. To better understand the intricate and superposing phenomena of this process, we exploit an original approach based on needle electron tomography, EXAFS and photoluminescence. Needle electron tomography surprisingly shows a heterogeneous distribution of Er content in the silicon thin pores that until now couldn't be revealed by the sole use of scanning electron microscopy compositional mapping. Besides, while showing that pore filling leads to enhanced photoluminescence emission, we demonstrate that the latter is originated from both erbium oxide and silicate. These results give a much deeper understanding of the photoluminescence origin down to nanoscale and could lead to novel approaches focused on noteworthy enhancement of Er-related photoluminescence in porous silicon.