PII Types & Categories
Understanding the full spectrum of personally identifiable information is essential for building comprehensive detection systems. PII ranges from obvious identifiers like names to subtle quasi-identifiers that can re-identify individuals when combined.
Direct Identifiers
Direct identifiers can uniquely identify an individual on their own without additional context:
| PII Type | Examples | Detection Difficulty |
|---|---|---|
| Full Name | John Smith, Maria Garcia | Medium - NER needed |
| SSN | 123-45-6789 | Easy - regex pattern |
| Passport Number | A12345678 | Medium - varies by country |
| Driver's License | D123-4567-8901 | Hard - format varies by state |
| Email Address | user@example.com | Easy - regex pattern |
| Phone Number | +1 (555) 123-4567 | Medium - many formats |
| Biometric Data | Fingerprints, facial geometry | Hard - binary data |
Quasi-Identifiers
Quasi-identifiers do not uniquely identify individuals alone, but can do so when combined. These are often overlooked in PII detection:
- Date of birth: Combined with zip code and gender, can re-identify 87% of U.S. residents
- Zip code / postal code: Narrows location significantly, especially in rural areas
- Gender: Reduces anonymity set by half
- Race / ethnicity: Further narrows identification in small populations
- Job title + employer: Often uniquely identifying for senior roles
- Medical conditions: Rare conditions combined with demographics can re-identify
Named Entity Types for NER-Based Detection
Named Entity Recognition (NER) models classify text spans into entity types. Key entity types for PII detection:
# spaCy default NER entity types relevant to PII PERSON # People, including fictional characters ORG # Companies, agencies, institutions GPE # Countries, cities, states (Geo-Political Entity) LOC # Non-GPE locations: mountain ranges, bodies of water DATE # Absolute or relative dates MONEY # Monetary values CARDINAL # Numerals that don't fall into other types # Extended PII entity types (custom models) EMAIL # Email addresses PHONE # Phone numbers SSN # Social Security Numbers CREDIT_CARD # Credit card numbers IP_ADDRESS # IPv4 and IPv6 addresses URL # Web addresses LICENSE # License plate or driver's license numbers MEDICAL_ID # Medical record numbers
Sensitive Data Categories
Regulations classify certain types of PII as especially sensitive, requiring elevated protection:
- Health information (PHI): Medical records, diagnoses, treatments, prescriptions, insurance IDs. Protected under HIPAA in the U.S.
- Financial information: Bank account numbers, credit card numbers, tax IDs, income data. Subject to PCI-DSS and financial regulations.
- Biometric data: Fingerprints, facial recognition templates, retinal scans, voiceprints. Increasingly regulated (BIPA in Illinois).
- Racial / ethnic data: Special category under GDPR Article 9. Requires explicit consent for processing.
- Political / religious beliefs: Protected under GDPR as special category data.
- Sexual orientation: Highly sensitive, protected under GDPR and various national laws.
HIPAA Safe Harbor De-identification
HIPAA defines 18 specific identifiers that must be removed for Safe Harbor de-identification:
- Names
- Geographic data smaller than a state
- Dates (except year) related to an individual
- Phone numbers
- Fax numbers
- Email addresses
- Social Security numbers
- Medical record numbers
- Health plan beneficiary numbers
- Account numbers
- Certificate/license numbers
- Vehicle identifiers and serial numbers
- Device identifiers and serial numbers
- Web URLs
- IP addresses
- Biometric identifiers
- Full-face photographs
- Any other unique identifying number or code
Ready to Go Deeper?
Live instructor-led courses from our partners. Affiliate disclosure.
AI & ML Courses - 30% Off
Live instructor-led AI, machine learning, data science, and cloud courses for working professionals. Use code Limited30 at checkout.
EdurekaDataCamp - AI & Data Science
Hands-on Python, machine learning, and AI courses with interactive exercises and real projects.
DataCampedX - Top AI Courses
University-level AI courses from MIT, Harvard, Stanford. Earn certificates that employers recognize.
edX