A potentially useful line of future work would be to accommodate holiday effects while estimating the density of electricity consumption, where each special day is modelled separately. Another potential area of research would be to investigate the incorporation of different clustering techniques, such as those considered by Dai and Kuosmanen (2014), to identify consumers with similar electricity consumption patterns. This information could potentially be useful for scaling the modelling framework to a large number of consumers, without incurring huge computational costs.