HemeraScope · EventsDemo
HemeraScope · Events

Reproducibility

Same inputs, same number. Every published footprint can be replayed bit-identically.

Reproducibility receipt

The published event footprint has been re-run against its frozen inputs hash and matches bit-identically. No upstream record has been silently mutated since publish.

Bit-identical · verified
Bit-identical to the original publication. Inputs hash 9f3a7c2d8e1b… · methodology event-v1.2.0

Inputs (frozen at publish)

Inputs hash (SHA-256)
9f3a7c2d8e1b4f6a52c8d9e0a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0
Publication slug
pawston-annual-2026-v1
Published at
2026-04-04 15:22:18 UTC
Methodology version
event-v1.2.0
Monte Carlo iterations
5,000
MC seed (first 16 hex)
a1b2c3d4e5f60718

Pinned emission factor sets

SourceVersionUsed forRole
DEFRA UK government GHG conversion factors2024 v1.1Travel, energy, wastePinned calculation source
Cornell Hotel Sustainability Benchmarking Index (CHSB)2023AccommodationBenchmark default for low-data estimates, carried at wide uncertainty
Poore & Nemecek (food)2018CateringBenchmark default for low-data estimates, carried at wide uncertainty
Exiobasev3.8.2Spend-based residualsPinned calculation source
Factor sets are pinned by version at publish. A new factor release does not silently update a published figure — re-running with a newer set creates a fresh publication with a different inputs hash. Benchmark-default sources back low-data estimates only and are carried at wide uncertainty.

What the inputs hash covers

The hash is a SHA-256 over a canonicalised JSON payload covering every input that feeds the calculator:

  • · Event metadata — start/end, attendee counts, gross floor area, gross revenue.
  • · Methodology version (event-v1.2.0).
  • · Activity records — every non-soft-deleted row's quantity, unit, factor id, factor version, category and scope, ordered by id.
  • · Allocation decisions — rule used, basis, justification.

Decimals serialise as strings (full precision preserved). Dict keys are sorted before hashing so payload-ordering can never shift the hash. Two publish runs with identical inputs produce the same hash; if anything moves, the diff is materialised in the methodology PDF change-log.

Replay result

Expected total
47,300 kgCO₂e
Recomputed total
47,300 kgCO₂e
Δ
0.000 kg
σ band width change
0.000 kg
Why this matters

The inputs hash is the spine of the publish workflow. An auditor looking at last year's report can ask HemeraScope to replay the same inputs and get the same number back. If anything drifted — a factor was retroactively edited, a record was silently mutated — the replay diverges and the diff is on the page.

Replay an individual line

The replay command isn't only for the headline. Every emission line in the atlas carries its own replay metadata. Open any line in the drawer — the replay command at the bottom regenerates that line's number from its pinned inputs.