Enhancing the spatio-temporal observability of distributed energy resources (DERs) is crucial for achieving secure and efficient operations in distribution grids. This paper puts forth a joint recovery framework for residential loads by leveraging the complimentary strengths of heterogeneous types of measurements. The proposed approaches integrate the low-resolution smart meter data collected for every load node with the fast-sampled feeder-level measurements provided by limited number of phasor measurement units. To address the lack of data, we exploit two key characteristics for the loads and DERs, namely the sparse changes due to infrequent activities of appliances and electric vehicles (EVs) and the locational dependence of solar photovoltaic (PV) generation. Accordingly, meaningful regularization terms are introduced to cast a convex load recovery problem, which will be further simplified to reduce computational complexity. The load recovery solutions can be utilized to identify the EV charging events at each load node and to infer the total behind-the-meter PV output. Numerical tests using real-world data have demonstrated the effectiveness of the proposed approaches in enhancing the visibility of these grid-edge DERs.