Table of Contents
This guide shows how to use metadata-grammar for practical data cartography—mapping reality through systematic observation and federation of phenomenal databases.
Description: Phenomenal databases we’ve created—observed, measured, mapped.
Properties: - Transparent and inspectable - Schema-defined structure - Known provenance - Registered with verisim
How to work with it: - Catalog databases using metadata-grammar - Express spatiotemporal coverage explicitly - Track quality and uncertainty - Link to other databases in federation
Description: Domains we know exist but haven’t measured.
Examples: - Geographic gaps (deep ocean, polar regions) - Temporal gaps (pre-satellite era, future projections) - Population gaps (underrepresented groups) - Variable gaps (unmeasured risk factors)
How to work with it: - Explicitly mark blind spots in metadata - Prioritize exploration efforts - Track coverage metrics over time - Guide data collection to fill gaps
Description: Domains we don’t even know to look for.
Properties: - Beyond current conceptual frameworks - Require new instruments or theories - True epistemic darkness
How to work with it: - Acknowledge in coverage estimates - Track discovery of new domains - Maintain epistemic humility - Document surprises when found
@prefix mg: <https://hyperpolymath.org/ns/metadata-grammar#> .
@prefix geo: <http://www.opengis.net/ont/geosparql#> .
@prefix time: <http://www.w3.org/2006/time#> .
:MyObservation a mg:PhenomenalDatabase ;
# What domain of reality does this map?
mg:observes mg-domain:UrbanAirQuality ;
# From what perspective?
mg:perspective "ground-sensor-network" ;
mg:methodology "PM2.5 measurements every 5 minutes" ;
mg:instruments :SensorArray-2025 .:MyObservation
# Spatial coverage (where is the white box?)
mg:spatialCoverage [
geo:hasGeometry "POLYGON((...))"^^geo:wktLiteral ;
mg:spatialResolution "100 meters" ;
rdfs:label "Greater London area"
] ;
# Temporal coverage (when is the white box?)
mg:temporalCoverage [
time:hasBeginning "2020-01-01T00:00:00Z"^^xsd:dateTime ;
time:hasEnd "2025-01-31T23:59:59Z"^^xsd:dateTime ;
mg:temporalResolution "5 minutes"
] .:MyObservation
# Spatial blind spots we're aware of
mg:knownBlindSpots [
mg:spatial [
rdfs:label "Indoor environments" ;
mg:reason "Sensors only outdoors"
] ;
mg:spatial [
rdfs:label "Private property" ;
mg:reason "Access restrictions"
] ;
# Temporal blind spots
mg:temporal [
rdfs:label "Pre-2020 period" ;
mg:reason "Sensor network not deployed"
] ;
# Variable blind spots
mg:variables [
rdfs:label "Ultrafine particles (<PM0.1)" ;
mg:reason "Sensor limitations"
]
] .:MyObservation
# Overall quality assessment
mg:qualityLevel mg:High ;
mg:uncertainty mg:Low ;
# Known biases
mg:knownBiases [
rdfs:label "Urban heat island effect on sensor readings" ;
mg:magnitude "±3% systematic error"
] ;
# Limitations
mg:limitations [
rdfs:label "Weather interference during heavy rain" ;
mg:impactedMeasurements 0.05 # 5% of measurements affected
] ;
# Validation status
mg:validatedAgainst :ReferenceStation-EPA ;
mg:validationCorrelation 0.95 .@prefix prov: <http://www.w3.org/ns/prov#> .
:MyObservation
# How was this created?
prov:wasGeneratedBy :AirQualityMonitoringProject ;
prov:wasAttributedTo :ResearchTeam-UCL ;
# When?
prov:generatedAtTime "2025-01-31T12:00:00Z"^^xsd:dateTime ;
# From what sources?
prov:wasDerivedFrom :RawSensorData-2020-2025 ;
# Using what methods?
prov:used :CalibrationProtocol-v2 ;
prov:used :QualityControlAlgorithm-v3 .@prefix vsim: <https://hyperpolymath.org/ns/verisim#> .
:MyObservation
# Join the federation
mg:registeredIn :GlobalAtlas ;
vsim:indexedBy :GlobalAtlas ;
# Enable temporal queries
vsim:stored-in :VerisimDBInstance ;
vsim:version-history :MyObservation-History ;
# Enable discovery
mg:accessEndpoint <https://api.example.org/air-quality> ;
mg:sparqlEndpoint <https://api.example.org/sparql> .Query verisim for coverage:
PREFIX mg: <https://hyperpolymath.org/ns/metadata-grammar#>
PREFIX geo: <http://www.opengis.net/ont/geosparql#>
# Find all databases covering London during 2020-2025
SELECT ?db ?title ?coverage
WHERE {
?db a mg:PhenomenalDatabase ;
mg:title ?title ;
mg:spatialCoverage ?spatial ;
mg:temporalCoverage ?temporal .
?spatial geo:sfIntersects :LondonPolygon .
?temporal time:hasBeginning ?start ;
time:hasEnd ?end .
FILTER(?start <= "2025-12-31"^^xsd:date)
FILTER(?end >= "2020-01-01"^^xsd:date)
}Find blind spots across the federation:
PREFIX mg: <https://hyperpolymath.org/ns/metadata-grammar#>
# What spatial regions are blind spots in air quality data?
SELECT ?region (COUNT(?db) as ?coverage)
WHERE {
?db a mg:PhenomenalDatabase ;
mg:observes mg-domain:AirQuality ;
mg:knownBlindSpots ?blindspot .
?blindspot mg:spatial ?region .
}
GROUP BY ?region
ORDER BY DESC(?coverage)Federated query across multiple databases:
PREFIX mg: <https://hyperpolymath.org/ns/metadata-grammar#>
# Correlate air quality with health outcomes
SELECT ?location ?pm25 ?asthmaRate
WHERE {
# Air quality database
SERVICE <https://airquality.example.org/sparql> {
?observation mg:location ?location ;
mg:pm25 ?pm25 .
}
# Health database
SERVICE <https://health.example.org/sparql> {
?outcome mg:location ?location ;
mg:asthmaRate ?asthmaRate .
}
}:AirQualityDB_v2020 a mg:PhenomenalDatabase ;
vsim:validTime "2020-01-01" ;
mg:spatialCoverage :CentralLondon ; # Limited coverage
mg:uncertainty mg:High ; # Early deployment
mg:knownBlindSpots [
mg:spatial :OuterLondon ; # Not yet covered
mg:variables :NO2 # Not yet measured
] .:AirQualityDB_v2023 a mg:PhenomenalDatabase ;
vsim:validTime "2023-01-01" ;
vsim:supersedes :AirQualityDB_v2020 ;
mg:spatialCoverage :GreaterLondon ; # Expanded!
mg:uncertainty mg:Medium ; # Improved calibration
# Blind spot filled
mg:cartographicDelta [
mg:blindSpotFilled :OuterLondon ;
mg:newVariable :NO2 # Now measured
] ;
# Remaining blind spots
mg:knownBlindSpots [
mg:variables :Ultrafines # Still not measured
] .:AirQualityDB_v2025 a mg:PhenomenalDatabase ;
vsim:validTime "2025-01-31" ;
vsim:supersedes :AirQualityDB_v2023 ;
mg:spatialCoverage :GreaterLondon ;
mg:spatialResolution "50 meters" ; # Improved from 100m
mg:uncertainty mg:Low ; # Mature deployment
mg:cartographicDelta [
mg:resolutionImproved "100m → 50m" ;
mg:uncertaintyReduced 0.15 ;
mg:newVariable :Ultrafines # Finally added!
] ;
# Minimal blind spots remaining
mg:knownBlindSpots [
mg:spatial :PrivateProperty # Irreducible
] ."What did we know about air quality on 2022-06-01?"
PREFIX vsim: <https://hyperpolymath.org/ns/verisim#>
PREFIX mg: <https://hyperpolymath.org/ns/metadata-grammar#>
SELECT ?coverage ?uncertainty ?blindSpots
WHERE {
# Query verisim at specific time
vsim:AtTime("2022-06-01") {
:AirQualityDB
mg:spatialCoverage ?coverage ;
mg:uncertainty ?uncertainty ;
mg:knownBlindSpots ?blindSpots .
}
}:GlobalAtlas a mg:CartographicIndex ;
# Overall coverage
mg:whiteBoxCoverage 0.15 ;
mg:knownDarknessCoverage 0.25 ;
mg:unknownDarknessCoverage 0.60 ;
# Federation size
mg:federationSize 1500000 ;
mg:totalObservations 5.2e15 ; # 5.2 quadrillion data points
# Domain coverage
mg:domainsCataloged [
mg-domain:Climate,
mg-domain:Economy,
mg-domain:Health,
mg-domain:Environment,
# ... 500+ domains
] ;
# Spatial coverage
mg:earthSurfaceCoverage 0.73 ; # 73% of Earth surface
mg:oceanCoverage 0.35 ; # 35% of ocean volume
mg:atmosphereCoverage 0.60 ; # 60% of atmosphere
# Temporal coverage
mg:historicalCoverage 0.20 ; # 20% of human history
mg:recentCoverage 0.95 ; # 95% of last decade
# Epistemic humility
mg:confidenceInCoverage mg:Low ; # Honest uncertainty
rdfs:comment "These estimates themselves have high uncertainty" .:GlobalAtlas
mg:explorationQueue [
(mg:priority 1) [
mg:domain mg-domain:DeepOceanBiology ;
mg:currentCoverage 0.05 ;
mg:expectedImpact mg:VeryHigh ;
mg:feasibility mg:Medium
] ;
(mg:priority 2) [
mg:domain mg-domain:SubSurfaceGeology ;
mg:currentCoverage 0.10 ;
mg:expectedImpact mg:High ;
mg:feasibility mg:Low
] ;
(mg:priority 3) [
mg:domain mg-domain:MicrobialDiversity ;
mg:currentCoverage 0.02 ;
mg:expectedImpact mg:High ;
mg:feasibility mg:High
]
] .Bad (pretends complete coverage):
:BadDatabase a mg:PhenomenalDatabase ;
mg:spatialCoverage :EntireWorld . # Overconfident!Good (honest about limitations):
:GoodDatabase a mg:PhenomenalDatabase ;
mg:spatialCoverage :MeasuredRegions ;
mg:knownBlindSpots [
mg:spatial :UnmeasuredRegions ;
mg:reason "Instrument limitations"
] ;
mg:estimatedUnknownBlindSpots mg:Moderate . # Epistemic humility:MyDatabase
mg:uncertainty mg:Medium ;
mg:uncertaintyEstimate [
mg:systematic "±5%" ;
mg:random "±2%" ;
mg:methodological "Unknown, possibly large"
] .:MyDatabase
mg:perspective "urban-sensor-network" ;
mg:knownBiases [
rdfs:label "Urban bias" ;
mg:description "Rural areas underrepresented" ;
mg:impactLevel mg:High
] ;
mg:samplingStrategy "Convenience sampling (non-random)" .:MyDatabase
# Register with atlas
mg:registeredIn :GlobalAtlas ;
# Declare relationships
mg:complementedBy :OtherDatabase ; # Fills your blind spots
mg:contradicts :OlderDatabase ; # Corrects errors
mg:derivedFrom :RawData ; # Provenance
# Enable discovery
mg:sparqlEndpoint <https://api.example.org/sparql> .:MyDatabase_v2
vsim:supersedes :MyDatabase_v1 ;
mg:cartographicDelta [
mg:improvement "Calibration protocol updated" ;
mg:blindSpotFilled :PreviousGap ;
mg:uncertaintyReduced 0.10
] ;
mg:whatChanged [
rdfs:comment "Discovered systematic error in v1" ;
rdfs:comment "Applied retrospective correction" ;
rdfs:comment "Added 500 new sensors in undersampled regions"
] .:GlobalClimateDB a mg:PhenomenalDatabase ;
mg:observes mg-domain:GlobalClimate ;
mg:spatialCoverage :GlobalLandAndOcean ;
mg:temporalCoverage "1850-2025" ;
mg:knownBlindSpots [
mg:spatial :DeepOcean ;
mg:spatial :PolarRegions-pre1950 ;
mg:temporal :pre-1850
] ;
mg:uncertainty [
mg:spatial :GlobalLand mg:Low ;
mg:spatial :Ocean mg:Medium ;
mg:temporal :recent mg:Low ;
mg:temporal :pre-1900 mg:High
] .:HumanGenomeDB a mg:PhenomenalDatabase ;
mg:observes mg-domain:HumanGenomicVariation ;
mg:populationCoverage :GlobalSample ;
mg:knownBiases [
rdfs:label "European ancestry overrepresentation" ;
mg:magnitude "80% of samples from European descent"
] ;
mg:knownBlindSpots [
mg:populations :IndigenousGroups ;
mg:variants :RareVariants # <0.1% frequency
] .:SocialBehaviorDB a mg:PhenomenalDatabase ;
mg:observes mg-domain:OnlineSocialBehavior ;
mg:dataSources :TwitterAPI, :FacebookAPI ;
mg:knownBiases [
rdfs:label "Platform bias" ;
mg:description "Only captures users of these platforms"
] ;
mg:knownBlindSpots [
mg:populations :NonInternetUsers ;
mg:populations :PrivateInteractions ;
mg:behaviors :OfflineSocialLife
] ;
mg:estimatedCoverageOfTotalPhenomenon 0.05 . # Only 5% of social behavior!metadata-grammar enables systematic cartography of reality through data:
-
Create phenomenal databases with explicit coverage boundaries
-
Mark known blind spots (Zone 2) honestly
-
Acknowledge unknown darkness (Zone 3) humbly
-
Register with verisim to join the federation
-
Track cartographic evolution as understanding improves
-
Guide exploration into darkness to expand the white box
We are cartographers of the digital age. The boxes are white. The territory is dark. Let’s explore.