Skip to documentation content

H00 — Start here (humans)

TUNES = Transit User Noise Experience Survey.

Open passenger programme for measuring noise on ordinary rail journeys — starting with the London Underground — and publishing both a quieter-journey map and a versioned open dataset with an honest methodology.

Not operated, approved, or endorsed by TfL.

Two doc layers:

LayerPrefix / folderForStyle
HumanH00H19Reading, decisions, onboardingCore briefs + linked topic briefs
Machinemachine/Implementation, tests, deep research, frequent editLong · detailed · linked ADRs

Why this exists

Professional surveys answer engineering and regulatory questions. TUNES adds repeated evidence from ordinary passenger journeys without replacing those surveys or claiming engineering-grade or health-risk authority.

Read the baseline rationale in the README and the deeper scientific and civic case in H19 — Why TUNES exists.

What TUNES does (and does not)

Does: iPhone-first journey measurement · optional perception survey (kept separate) · map with uncertainty/sample size · open methodology + versioned dataset with provenance and quality tiers.

Does not: replace TfL / Class 1 surveys · imply TfL endorsement · claim health risk or engineering-grade accuracy at foundation · silently upload raw audio or collapse into one unexplained “score”.

How the programme works

Passenger records on iPhone
        ↓
Local processing + journey alignment (user reviews sections)
        ↓
Consent screen shows exactly what will leave the device
        ↓
Derived metrics upload (raw audio stays on device by default)
        ↓
Server inspect → validate → quality-flag → aggregate
        ↓
Versioned open release + public map

Comparison is usually station-to-station section + duration. Privacy default: record locally → review → consent → upload derived features only.

The three repositories

Not a monorepo — science, capture, and public site change at different speeds.

workspace/
  tunes/          ← you are here (programme + open data)
  tunes-ios/      ← recorder app
  tunes-web/      ← landing, map, pipeline
RepoOwnsDoes not own
tunesCharter, governance, ADRs, research, methodology, schemas, licences, claim language, open releasesApp UI, live map runtime
tunes-iosPCM/motion capture, offline local processing, alignment UX, upload client, consent previewCanonical dataset / charter
tunes-webLanding, map UI, upload intake, pipeline jobsNative capture; charter / methodology source of truth
tunes-ios  --derived upload-->  tunes-web  --validated release-->  tunes
tunes      --schema / methodology / licences-->  tunes-ios + tunes-web

Pilot posture

London-first; dense repeated observations over shallow network-wide coverage. Generic railway data model so other cities/modes can be instances later. iPhone-first; Android / wearables / external mics are later expansions.

Core reading order (≈15 min)

#DocOne line
H01What is TUNESMission + what we will / won’t claim
H02Repostunes · tunes-ios · tunes-web
H03DecisionsAccepted ADRs at a glance
H04How it worksCapture → review → pipeline → open data
H05PrivacyDerived-only default; raw stays on device
H06QualityTiers A–E · survey-grade honesty
H07RoadmapPhases 0–10
H08PilotLondon Underground-first
H09RisksTop risks only
H10GlossaryShared terms
H11Legal & governanceEntity, legal pack, DPIA / beta gates

Topic briefs

Use these as references after the core path.

#DocOne line
H12Measurement philosophyWhat is measured; repeatability and uncertainty
H13Public mapAggregation, confidence, colour and privacy
H14RecorderMobile capture, alignment, metadata and upload
H15Railway noise taxonomyProvisional source labels and detectability limits
H16SchemaWhy the data entities and relationships exist
H17PortabilityGeneric core versus network configuration
H18AssumptionsWorking assumptions and validation needs
H19Why TUNES existsPhilosophical and technical centre

Also at repo root: README (programme baseline and repository overview).

Machine index

machine/00-index-full.md · machine/README.md