The Taxonomy

A shared vocabulary for describing technology.

Each category is a question people ask about a technology, and its elements are the answers. Choose a category, then an element, to see its icon and what it means. New to the terms? Read the concepts.

Output Data

Elements 10

What does this system produce?

The output data this deployed system produces when it operates — decisions, content, and actions.

About this category

Output Data

What does this system produce?

The output data this deployed system produces when it operates — decisions, content, and actions.

Datachain

Answer: Some systems only collect. Answer it when the system produces decisions, content or actions of its own.
Icon shape: Data in motion: what goes in, what processes it, and what comes out.

Context: Personal Information

How identifiable the data is when this system produces it. The element name says what the output is about; this colour says how directly it identifies a person.

Anonymized dataOutput about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so the result can’t reasonably be tied back to an individual.

Pseudonymous dataThe output is labelled with a token (hash, ID, template) that lets this system tie the result to the same person across events without revealing a name. Reidentification is possible with extra information.

Identifiable dataThe output names a person directly (name, address, account, recognisable face or voice, plate) or carries a token this system resolves to legal identity in the same step.

Variables

Description Required

Data about a measurementThe system produces numeric measurements over time — sensor outputs, derived quality scores, forecasted values. Examples: an hourly air-quality score for a plaza, a predicted ridership curve for tomorrow, an aggregated noise-level reading published to a public dashboard.

Data about a placeThe system produces location, route, region, or floor-plan output. Examples: a wayfinding route shown on a kiosk, a heatmap of foot-traffic density, a geofence boundary published to downstream systems.

Data about behaviourThe system produces records of what people did — paths walked, items selected, dwell times, interaction patterns. Examples: a weekly report of most-visited destinations, a session log of touchscreen interactions, an inferred route someone took through a venue.

Biometric dataThe system produces output containing biological signals from a person's body — face templates, voice embeddings, gait vectors, gesture sequences. Examples: an enrolled face template stored downstream, a voice embedding for later matching, a gait signature exported to another system.

A decision about youThe system produces a binding determination about a person — eligibility, classification, ranking, or yes/no. Examples: an access-control system deciding to open a gate, an automated benefits eligibility result, an automated screening that rules someone out.

Generated contentThe system produces new content — text, images, audio, or video that did not exist before. Examples: an LLM-written service advisory shown on a sign, a synthesized voice announcement played over a public-address system, a generated illustration on a digital kiosk.

Operational dataThe system produces administrative output — schedules, routes, occupancy estimates, budget allocations, or other records about how a place or service runs. Not about any one person. Examples: an optimized trash-collection route, a school occupancy projection, a recommended budget allocation for next quarter.

A physical actionThe system triggers something in the physical world — a door unlocking, a light turning on, an alert sounding, signage changing, HVAC adjusting. Examples: a turnstile that opens on facial match, a public-address alert played automatically, an HVAC adjustment commanded by a building model.

A recommendation or predictionThe system produces an advisory output — a suggestion, forecast, risk score, or ranking. Distinct from a binding decision because it advises rather than determines. Examples: a signage system suggesting an alternate route based on predicted crowding, a forecasted demand curve for tomorrow, a recommended next destination on a kiosk.

Sensitive personal dataThe system produces output about health, finances, beliefs, sexuality, immigration status, or other categories that carry legal and social risk. Examples: a triage priority assigned at a clinic, an eligibility result sent to a benefits portal, an inferred protected attribute exported downstream.

Sensitive personal data

The system produces output about health, finances, beliefs, sexuality, immigration status, or other categories that carry legal and social risk. Examples: a triage priority assigned at a clinic, an eligibility result sent to a benefits portal, an inferred protected attribute exported downstream.

Context: Personal Information

How identifiable the data is when this system produces it. The element name says what the output is about; this colour says how directly it identifies a person. Choose a value to see its icon.

Variables

Description Required

Sources

GDPR Article 9 — Special categories of personal dataRegulation (EU) 2016/679, Article 9.
Apple App Store privacy labels — Sensitive Info / Health & Fitness / Financial InfoApple.