Process Integration techniques are powerful in analysing the energy usage of industrial plants and designing retrofit solutions for increasing their energy efficiency. However they are far from constituting the industrial practice. One primary cause for this is the severe time consumption they entail, especially in acquiring the necessary process data. A recently developed method, the Required Data Reduction Analysis (RDRA), aims at reducing this time consumption. It makes use of uncertainty and sensitivity analysis techniques, together with optimisation tools, in order to systematically identify: (i) a reduced number of process parameters, whose information need to be acquired with high detail, and (ii) the maximum acceptable uncertainty of each of them. The paper presents recent additions to the method, expanding its capabilities and addressing two open questions highlighted in previous work. In particular: (i) three sensitivity analysis techniques are compared, based on computational effort and precision, (ii) a novel algorithm for minimising the number of parameters needing detailed data acquisition is proposed and (iii) the impact of the uncertainties characterisation on the outcome of the analysis is assessed. The results of the analysis considering four different case studies provided evidence that the Standardised Regression Coefficient method for sensitivity analysis should be preferred in the Required Data Reduction Analysis, and that the uncertainty characterisation did not jeopardise the results. Moreover, the proposed minimisation algorithm proved to be effective and fast in reducing the number of parameters. A reduction of 5 to 13 parameters out of 19 to 22 was achieved on the four case studies by requiring 14000 to 22000 model simulations.