SIGDG Ontology

The Signals Data Governance ontology (prefix SIGDG) is grounded in BFO 2020 (Basic Formal Ontology). BFO provides the upper-level categories; SIGDG specializes them for data governance.

BFO Grounding

BFO:0000031 generically dependent continuant
  └── SIGDG:0001 InformationEntity         ── "data that can be stored and transferred"

BFO:0000019 quality
  └── SIGDG:1001 SensitivityLevel          ── "inheres in an information entity"

BFO:0000023 role
  └── SIGDG:2001 DataSubjectRole           ── "externally grounded in data governance context"

BFO:0000015 process
  └── SIGDG:3001 DataLifecycleProcess      ── "transformation, migration, classification"

BFO’s generically dependent continuant (GDC) is the natural home for information entities — they depend on some material carrier (disk, memory) but can be copied and transferred between carriers. Sensitivity levels are qualities that inhere in information entities. Data subject roles are roles played by persons in a governance context.

CURIE Prefix

SIGDG → https://signals360.dev/ontology/dg/
BFO   → http://purl.obolibrary.org/obo/BFO_

Information Entity Hierarchy

SIGDG:0001  InformationEntity               (⊑ BFO:0000031)
├── SIGDG:0010  IdentityInformation         "data that identifies a natural or legal person"
│   ├── SIGDG:0011  GovernmentIdentifier    "state-issued ID: passport, SSN, driver's license"
│   ├── SIGDG:0012  PlatformIdentifier      "account-scoped ID: username, email, employee ID"
│   └── SIGDG:0013  DeviceIdentifier        "hardware-bound ID: IMEI, MAC address, UDID"
├── SIGDG:0020  PersonalInformation         "data about a natural person"
│   ├── SIGDG:0021  FinancialInformation    "monetary data: balances, scores, compensation"
│   │   └── SIGDG:0022  PaymentInformation  "payment instruments: card numbers, bank accounts"
│   ├── SIGDG:0023  DemographicInformation  "attributes: age, gender, ethnicity, education"
│   ├── SIGDG:0024  HealthInformation       "medical conditions, biometrics, genetic data"
│   ├── SIGDG:0025  ContactInformation      "addresses, phone numbers, email"
│   └── SIGDG:0026  CredentialInformation   "passwords, PINs, keys, security questions"
├── SIGDG:0030  BusinessInformation         "data about an enterprise"
│   ├── SIGDG:0031  ProprietaryInformation  "trade secrets, technical documentation"
│   └── SIGDG:0032  RegulatoryInformation   "SEC filings, compliance records"
├── SIGDG:0040  SystemInformation           "infrastructure and configuration data"
│   ├── SIGDG:0041  ConfigurationData       "IPs, cluster membership, file paths"
│   └── SIGDG:0042  AccessControlData       "permissions, LDAP groups, session state"
├── SIGDG:0050  TransactionInformation      "event records: orders, sessions, timestamps"
└── SIGDG:0060  TransformationMetadata      "provenance: encryption, masking, hashing records"

Sensitivity Levels

Sensitivity is modeled as a BFO quality (BFO:0000019) that inheres in an information entity:

CURIE	Label	Definition
`SIGDG:1010`	Public	Intended for unrestricted disclosure
`SIGDG:1020`	Internal	Not public but no harm if disclosed internally
`SIGDG:1030`	Confidential	Disclosure causes measurable harm to individuals or enterprise
`SIGDG:1040`	Restricted	Disclosure causes severe harm; regulated by law (GDPR, HIPAA, PCI)

Data Subject Roles

Data subjects are modeled as BFO roles (BFO:0000023), allowing the same information entity to carry different sensitivity depending on whose data it is:

CURIE	Label	Example
`SIGDG:2010`	CustomerRole	End-user, subscriber, app developer
`SIGDG:2020`	EmployeeRole	Employee, contractor, vendor
`SIGDG:2030`	EnterpriseRole	Corporate entity, partner organization

A column tagged SIGDG:0025 ContactInformation with subject SIGDG:2010 CustomerRole may be Restricted, while the same column with subject SIGDG:2020 EmployeeRole may be Confidential.

Data Lifecycle Processes

Classification itself is a BFO process (BFO:0000015):

CURIE	Label	Definition
`SIGDG:3010`	ClassificationProcess	Assigning SIGDG categories to an entity
`SIGDG:3020`	MigrationProcess	Moving data between storage tiers
`SIGDG:3030`	MaskingProcess	Applying de-identification transforms

Training Data

The classification model is trained on synthetic datasets spanning the SIGDG hierarchy:

Table	SIGDG Category	Columns	Rows
`identity_data`	SIGDG:0010 IdentityInformation	59	100
`personal_data`	SIGDG:0020 PersonalInformation	215	100
`business_data`	SIGDG:0030 BusinessInformation	21	100
`system_data`	SIGDG:0040 SystemInformation	43	100
`transaction_data`	SIGDG:0050 TransactionInformation	17	100
`metadata_info`	SIGDG:0060 TransformationMetadata	77	100
`annotations`	(ontology reference)	12	10

Each table’s columns serve as positive examples for their SIGDG category. The column names themselves (e.g., drivers_license_number, payment_card_number, trade_secrets) are the features the model learns to classify.

Keyboard shortcuts

Signals