Classification of Building Properties from the German Census Data for Energy Analysis Purposes

Abstract

The building sector is an important target for reducing urban energy consumption. Detailed data on the building stock is needed for modelling urban building energy demands but its availability is often insufficient. In Germany, the largest available public database about the building stock is the census national database, containing critical attributes for building characterization on a national scale, such as building age, construction type, and number of residents. However, the detailed information about the individual buildings is restricted by national data privacy laws and the information is found in an aggregated format. This study shows statistical and machine learning approaches to take the census data and disaggregate its information to each individual building in order to use that information for urban building energy modelling. This study presents a classification model for the following parameters of the 2011 census: building age, building form and heating type, and Number of residents. A study case was conducted in Oldenburg, Germany.