AS/NZS 5125.1 Report Analyser | Knowledge Base

Overview

A tool that accepts an uploaded AS/NZS 5125.1 test report, extracts the performance data, and produces a suite of visualisations and exportable data tables. The core output is a COP performance curve for the uploaded product overlaid against the full market range, plus a distribution chart showing where the product sits relative to all 5125.1-tested products in the EnergyAE database. This feature has dual purpose: internal use for compliance work and client-facing use for manufacturers who want to understand and market their product’s performance position. The visual outputs must therefore be polished enough to drop into a marketing document or client presentation.

User Stories

As an EnergyAE consultant, I want to upload a 5125.1 report and immediately see structured performance data without manually extracting it from a PDF so I can work faster. As a manufacturer, I want to see how my product’s COP compares to the rest of the market so I can understand its competitive position. As a marketing team member, I want to export clean charts and data tables showing my product’s performance so I can use them in product literature and presentations. As Alastair, I want every 5125.1 report processed through this tool to automatically enrich the market database so the industry comparison visuals improve over time without manual effort.

Inputs

Report upload:

User uploads a single AS/NZS 5125.1 test report in PDF format System extracts structured performance data automatically (see extraction requirements below) User can review and correct extracted data before visualisations are generated Extracted data stored against a product record in the database

Manual input fallback

If extraction fails or produces errors, user can enter data manually via a structured form Form fields mirror the extracted data schema Manual entry should not feel like a punishment - clean tabular input, not a wall of fields

Data Extraction Requirements

The extraction pipeline must parse the following from a standard 5125.1 report. the developer to review a sample report before building to understand the document structure. Product identification:

Product name and model number Manufacturer / brand Test laboratory name Test report number Test date / report date Standard version tested against

Performance test points

5125.1 tests at multiple ambient temperature conditions. For each test condition the following must be extracted:

Ambient air temperature (°C) Rated heating capacity (kW) Power input (kW) COP (coefficient of performance) Hot water delivery volume (litres) where reported Test duration where reported First hour rating where reported

Extraction approach

Primary method: LLM-based extraction (pass PDF pages to Claude, return structured JSON) the developer to assess whether direct PDF text extraction is sufficient or whether vision-based extraction is needed for scanned/image PDFs Extracted data presented to user for confirmation before being committed to the database Confidence flag on each extracted field: high (cleanly parsed) vs low (uncertain, needs review)

Visualisations

COP vs Ambient Temperature Curve The primary chart. Shows the uploaded product’s COP performance across the ambient temperature range tested. Overlaid market range band:

Shaded band showing the full range of COP performance across all products in the market database (min to max at each temperature point) Secondary shaded band or line showing the 25th-75th percentile range (interquartile range) Median market COP line Uploaded product’s COP curve clearly differentiated (bold line, distinct colour)

Chart requirements

X-axis: ambient temperature (°C) Y-axis: COP Hover tooltip: exact COP value for the product and market median at any temperature point Legend clearly labelling: this product / market range / market IQR / market median Chart title includes product name and model Export as PNG and SVG

Heating Capacity vs Ambient Temperature Curve Same structure as COP chart but showing rated heating capacity (kW) on the y-axis. Market range band overlaid as above.
Power Input vs Ambient Temperature Curve Same structure, showing electrical power input (kW). Market range band overlaid.
Market Position Distribution Chart A distribution curve (kernel density estimate or histogram, the developer to propose) showing where all products in the database sit for a selected metric and temperature point.

User selects: metric (COP / capacity / power input) and temperature point (e.g. 20°C ambient) Distribution of all market products shown as a curve or histogram Uploaded product’s value marked as a vertical line with label Percentile position shown (e.g. “Your product is in the 78th percentile for COP at 20°C ambient”) This chart will improve in value as the database grows - make this visible to the user (e.g. “Based on N products in the database”)

Summary Performance Table Clean tabular display of all extracted test points with calculated fields. Intended for direct export. Columns:

Ambient temp (°C) Heating capacity (kW) Power input (kW) COP Performance vs market median COP (% above or below) Market percentile for COP at that temperature

Export Requirements

All exports must be production quality - these will go into manufacturer marketing materials and client presentations. Chart exports:

PNG at minimum 300 DPI SVG for vector quality Charts exported with clean white background, no platform UI chrome Optional: branded export with EnergyAE logo in corner (toggle)

Data table exports

Excel (.xlsx): formatted table with product details header, performance data, and calculated fields CSV: raw data for further processing

Full report export

Single PDF containing all charts plus summary table Product name, model, test report number, and generation date shown on each page Clean layout suitable for appending to a client deliverable the developer to review the docx/PDF skill before building this output

Market Database

This feature depends on a growing database of 5125.1 performance data. The comparison visuals are only meaningful with sufficient data points. Database population:

Every report processed through this tool automatically contributes its data to the market database (with user confirmation) Alastair to seed the database with historical 5125.1 data at launch - format to be agreed with the developer Each product record: brand, model, test date, all extracted performance points, source report reference

Data quality

Duplicate detection: if the same model number already exists in the database, prompt the user to confirm whether this is a new test or duplicate Version handling: if a product has multiple test reports (different dates), store all versions, use most recent for comparisons by default Database size shown to users on the distribution chart to set expectations on statistical reliability

Access and privacy

Product-level data (which brand has which COP) should not be visible to manufacturer clients viewing competitors Aggregated market range and distribution data is shown without identifying individual products EnergyAE staff can see full database including product identities Manufacturer clients see only their own product data plus anonymised market comparisons

UI / UX Direction

Upload step is prominent and simple: drag-and-drop zone or file picker, PDF only Extraction review step: clean table showing extracted fields with edit-in-place capability, confidence flags highlighted for low-confidence fields Results page: charts displayed in a logical sequence (COP curve first as primary output), summary table below Chart controls (temperature point selector for distribution chart) should update the chart instantly without a page reload Export buttons clearly grouped and labelled Mobile: upload and review usable on mobile; charts may be desktop-only for v1 given complexity the developer to produce wireframe before building

Out of Scope (v1)

Batch upload of multiple reports Automated report ingestion from laboratory portals (CVC, Intertek) Side-by-side comparison of two uploaded products AS/NZS 4234 simulation integration (linking test data to simulation inputs automatically) White-label export for manufacturer branding Public product performance database / directory

Data Model (indicative)

Products table:

product_id brand model_number manufacturer test_lab report_number test_date uploaded_by uploaded_at source_report_url (stored PDF reference) is_market_db_contributor (boolean - user confirmed contribution)

Performance points table

point_id product_id (foreign key) ambient_temp_c heating_capacity_kw power_input_kw cop delivery_volume_l (nullable) first_hour_rating (nullable) extraction_confidence (high / low / manual)

Acceptance Criteria

The feature is done when:

Open Questions (resolve before build starts)

Alastair to provide 3 sample 5125.1 reports before the developer begins extraction work. These are needed to understand document structure and assess whether text extraction or vision-based extraction is required. What is the seed dataset size for the market database at launch? The distribution chart needs a minimum meaningful number of products (suggest at least 15-20) to be useful at launch. Alastair to assess what historical data is available. Should manufacturer clients be able to use this tool to upload their own reports, or is it EnergyAE staff only in v1? The branded export option (EnergyAE logo on charts) - is this always on, always off, or a user toggle? For the distribution chart, the developer to recommend kernel density estimate vs histogram based on expected dataset size. With a small database a histogram may be misleading; KDE handles small N more gracefully. Data privacy: are there any confidentiality obligations around test report data from CVC or Intertek that would prevent contributing specific product data to the market database? Alastair to confirm before the database contribution mechanic is built.

A Note on the Market Database

This feature compounds in value over time. Every report processed makes the comparison visuals more statistically meaningful. At launch with a small seed dataset the percentile claims will be directionally useful but not statistically robust - this should be communicated transparently to users (hence the “based on N products” label). Within 12-18 months of operation, given your volume of 5125.1 work, the database should become genuinely differentiated market intelligence that no competitor can easily replicate. That’s worth flagging to the developer as context: he’s not just building a chart tool, he’s building a proprietary dataset.