A computational analysis of Ireland’s national building stock at the intersection of energy performance and architectural heritage.
This paper cross-references two large-scale government datasets — the SEAI Building Energy Rating database (1 million+ records, 211 columns) and the National Inventory of Architectural Heritage (55,000 expert-catalogued records) — to produce a corrected model of Ireland’s historic building stock and its energy performance.
A critical statistical anomaly was identified in the BER data: a massive artificial spike of properties defaulted to a construction year of 1900, caused by surveyors lacking the expertise to date historic structures. Using the NIAH’s faithful decade-by-decade distribution, this spike was redistributed to produce a ‘healed model’ — revealing that protected heritage properties represent the overwhelming majority of the worst-performing G-rated housing stock nationally.
The paper argues that the recast EU Energy Performance of Buildings Directive cannot be implemented effectively without conservation-led strategies, and proposes a pro-rata grant model scaled to actual energy savings rather than the current flat-rate SEAI grant — providing the financial bridge that heritage retrofits require.
Presented nationally (May 2024) and internationally (Estonia, December 2025). Research contributed Irish data to the ICOMOS ISCES sub-committee on the EPBD recast, published January 2026.
Enter your details below and you’ll receive an immediate link to the full PDF.
Your details are used only to understand who is engaging with this research. No marketing emails, no third parties. To request removal email remove@architecturalintelligence.ie and your details will be deleted within 48 hours.
The paper is ready to download. If you have questions or would like to discuss the research, feel free to get in touch.
⬇ Download PDFOr email the author directly.
This paper is also intended as a showcase and demonstration of the technologies involved — Python, the Pandas library, and Jupyter Notebooks. Just as architects have always used a black notebook to record observations, sketches and calculations, the Jupyter environment is a digital equivalent where notes, text, code and visualisations coexist. The difference is the scale of interrogation it makes possible.
If you’re an architect who saw something in this paper that made you think “I could do this for my own practice or research” — register your interest below. If there is sufficient demand, a structured introduction to these tools, designed specifically for architects, will follow.
No commitment, no fee. Just a way to gauge interest. You’ll hear from david@architecturalintelligence.ie if a course runs. To be removed at any time email remove@architecturalintelligence.ie.
Thank you — interest noted.
If enough architects express interest, you’ll hear from david@architecturalintelligence.ie with details of dates and format. In the meantime, the Jupyter website is a good place to start exploring.