Industrial applications typically necessitate the in-teraction of heterogeneous software components, which makesthe design of an integrated system a demanding task. Specifically,although Matlab® and Java are among the most commonly usedprogramming languages in industrial practice, with each oneoffering its own advantages, however, their integration for real-time code prototyping is not straightforward. Motivated by thisproblem, this work proposes an efficient method based on theuse of sockets to integrate Matlab and Java code for designinga data processing platform tailored to smart water sensornetworks scenarios. The performance of the proposed approachis evaluated on two distinct tasks, namely, the recovery of missingvalues and the temporal super-resolution from streaming data.Experimental evaluation with real pressure data reveals thesuperiority of our methodology, in terms of reduced executiontimes, when compared against two well-established alternatives,namely, the use of standalone applications using input-outputfiles for executing Matlab code in Java-based environmentsand socket-based solutions implemented directly in a Matlabenvironment.