Validation approaches under GDPR constraints for bottom-up building stock energy data: Case Vienna

Abstract

The “Spatial Energy Planning” project, funded in the frame of the Austrian Energy Model Region “Green-Energy-Lab”, provides a detailed structured methodology for developing high granularity datasets focusing on heating transition. The methodology subjects input datasets from sources such as building energy performance certification, heating database, local district heating networks and statistical data through a series of spatial and statistical filtering and processing. The approach results in detailed building stock information on the spatial distribution, physical and energy properties, the end-use, and energy demand. However, options for validating the novel approach are limited due to the General Data Protection Regulation (GDPR) restrictions. Hence, in this paper, we developed a methodology and structure to validate the current outputs, with potential for application for datasets generated for other regions. The validation of the datasets was conducted on two levels. First, the allocation of the building’s geospatial attributes was checked, followed by a series of cross-validation measures and plausibility checks. For output datasets of each of the modules of the methodology, the distribution and statistical measures of the buildings based on end-use, construction period, physical properties, and energy consumption were analyzed. This facilitated the identification of possible outliers and their cause. If outliers were identified as errors originating from the methodology, potential alterations were proposed. This iterative approach provided a measure for improving the final results. The majority of the outliers were narrowed down to originate from the input datasets, thus emphasizing the importance of quality data inventory setup. In the second step, a comparative validation approach was undertaken. The outputs were tallied with the reference data acquired from the local distribution grid operator (Wiener Netze) on an aggregated level due to restrictions imposed by GDPR. Nevertheless, this allowed the evaluation of the data against measured data, which provided a solid basis for validation. Further detailed comparative analysis on energy-specific indicators would provide ground for conclusive validation of the methodology.