A new paper introduces the first automated data-driven validation of an EPC dataset
A new research paper by Kumar Raushan of the Irish Building Stock Observatory and the Built Environment Research and Innovation Centre (BERIC) in Technological University Dublin, funded by MaREI, introduces the first automated, data-driven validation of an EPC dataset. It adapts as the database evolves and scripts will be made available on the IBSO platform.
Seventeen unique filters were developed, revealing that 30% of EPC entries were erroneous and/or outlier data, with a staggering 80% related to misassessment of geometrical features. Errors in one field often correlated with errors in others, indicating some Assessors’ low responsibility towards data quality.
Upstream input data validation measures to our EPC system, need to be introduced by Sustainable Energy Authority of Ireland (SEAI) to reduce very avoidable Assessor mistakes when inputting data.
Paper Highlights
- A generalisable methodology for validation of a large EPC database is presented.
- Method democratises access to robust housing stock EPC data for all stakeholders.
- Data-driven methodology adapts to ever evolving EPC databases.
- Quantifies significant levels of erroneous data in a large EPC database.
- Makes recommendations to improve EPC data quality.
Learn more.
