Model Data From an EPC

An EPC contains a summary of a model, not the full model itself. The public certificate shows the rating, the band, estimated costs and emissions under standard conditions, and a list of recommendations, while the underlying input data, such as wall types, areas, glazing details and system specifications, is held in lodged data that is accessible through register and scheme routes for those with a legitimate purpose. Using EPC data for further modelling therefore means starting from a summary plus, where available, its inputs, and understanding that some inputs will be defaults rather than observations.

That distinction determines how far EPC data can be taken. Where input data is available, a modeller can rebuild a baseline model, test retrofit options against it, and identify which inputs drive the rating. Where only the public certificate is available, a modeller must infer inputs from the rating and recommendations, which is far weaker, because many input combinations can produce the same rating. A retrofit model built from a public certificate alone should be treated as a screening estimate, not as a design baseline.

Input quality is the main limit. EPC inputs include observed items, documentary items and defaults, and the certificate alone does not always distinguish them. A wall recorded as insulated may have been observed, documented or defaulted, and each basis carries different confidence. A modeller reusing EPC data should therefore grade inputs by provenance where possible, verify the influential ones against the building, and test sensitivity to the uncertain ones, rather than treating every EPC input as a measured fact.

Age is a further limit. An EPC up to ten years old may describe a home that has since been extended, insulated, re-glazed or given a new heating system, and reusing its data without checking for intervening work will model a home that no longer exists. A modeller should always ask what has changed since the assessment, and should commission a fresh assessment where changes are significant, rather than patching old data with guesses.

Before reusing EPC data, obtain the fullest data available through the legitimate route for your role, whether owner, adviser or professional, and check its date, method version and register reference. Verify the influential inputs against the building: heating system, controls, glazing, wall type and recorded insulation. Note which inputs rest on defaults, and record any intervening work since the assessment.

Also clarify the purpose. For early options screening, EPC data may be sufficient, with its limits stated. For detailed design, costing or compliance, a fresh survey is usually needed, because design decisions should not rest on defaulted inputs from an old assessment. Matching data quality to decision stakes is the core judgement in reusing EPC data.

By place, EPC data access differs. England and Wales, Scotland and Northern Ireland have their own register routes and data access rules, and the level of input detail available to owners, professionals and the public varies. EU certificate data follows national register rules under the EPBD framework, with varying openness. A modeller working across nations should check the access route for each property's system rather than assuming one route covers all.

Privacy also constrains reuse. Certificate data concerns a home and its occupants, and it should be used for the legitimate purpose that justifies access, such as advice for the owner, a transaction, or research on aggregated data, not circulated or repurposed casually. Aggregated reuse for policy or research should use anonymised or aggregated data where available.

A common mistake is to treat an EPC rating as a complete model, and to build retrofit advice on a rating without inputs. Another mistake is to reuse old EPC data after major work, producing a baseline that neither describes the current home nor supports a credible options comparison.

A further mistake is to copy EPC recommendations into a retrofit plan without checking suitability. Recommendations are generated from inputs, and where inputs are defaulted or outdated, recommendations may be unsuitable, already implemented, or impossible for the building. Each recommendation should be checked before it is adopted.

Keep the EPC, its input data where obtained, the verification notes and records of intervening work together as the modelling baseline file. That file lets a later modeller see exactly what came from the EPC and what was verified or updated.

Questions to ask include: what input data is available beyond the public certificate, how old is it, and which inputs have been verified. Those questions define what EPC data can support.