WORD STUDY

privacy

Privacy is both something people look for and something institutions turn into policies, controls, and rights.

71%

of the verified public sources on this page connect privacy to more than one system.

36%governance attention / 2025

42% idea
36% governance
22% pressure

  • WHAT PEOPLE OPEN
  • WHAT POLICIES SAY
  • WHERE RIGHTS AND RISKS MEET

01 / ATTENTION

What people opened changed.

20182025

+21.01percentage points since 2018

18
19
20
21
22
23
24
25
CONCEPTGOVERNANCEPRESSURE

02 / TEN WINDOWS

The selected topics do not move together.

“Data privacy” is a smaller window in this record. It contributes 0.52% of the selected ten-page attention total, while its 2025 traffic remains 47% of its own 2018 peak. The next group moves into policy, rights, and surveillance.

The final three windows narrow the view to state law, breach pressure, and digital rights. Every percentage still uses the same ten-page total.

GDPR / DIFFERENT TIME WINDOW

The frozen series starts in 2022, so it is not inserted into the shared 2018–2025 comparison. Its absence here is not a zero.

03 / POLICY LANGUAGE

Privacy becomes interface language.

Five complete frozen documents use different operational vocabularies.

The first two documents come from large online services. The last three come from a smaller service, a state regulator, and European law. Together they show how privacy language changes with the job each document has to do.

04 / PUBLIC SYSTEMS

One source can belong to several privacy systems.

DATA71%
RISK29%
CONSENT29%
PLATFORM29%
SECURITY18%
AI12%
RIGHTS12%

Each route uses the same 17-source denominator. A source can belong to several routes, so these percentages overlap rather than add to 100%.

17 VERIFIED SOURCE RECORDS

05 / WHAT MADE THE CUT

Missing evidence stays missing.

27.3%USED ON THIS PAGE

3 / 11 layers. Complete layers that support the visible findings.

18.2%MISSING OR TOO THIN

2 / 11 layers. Recorded as missing evidence, never converted to zero.

18.2%NOT COMPARABLE

2 / 11 layers. Real records whose units cannot share this page’s scales.

36.4%LEFT OUT

4 / 11 layers. Failed, partial, or outside the approved question.

Only three audited layers support the visible figures. The rest stay named here so a missing, partial, or incompatible source cannot silently become evidence—or be mistaken for zero.

SEE EVERY EVIDENCE STATE
USED HEREattention metrics

Ten frozen yearly pageview series have an explicit source boundary and complete 2018–2025 window.

UNAVAILABLEattention gdpr topic

The frozen GDPR page series begins in 2022 and cannot enter the 2018–2025 common-window comparison.

USED HEREplatform policy corpus

Five frozen documents pass the 1,000-token content floor.

USED HEREmodern institutional anchors

Seventeen high-confidence, non-manual-review anchors have reachable 2xx sources.

NOT COMPARABLEfrequency terms

Variable Ngram aliases and no fixed-release raw denominator.

NOT COMPARABLEgeo attention map

Academic production and news/source geography are mixed.

OUTSIDE THIS PAGEgeo elevation distribution

Elevation has no supported explanatory relationship to privacy attention.

UNAVAILABLEetymology first use

Direct first-use evidence remains unverified.

NOT SEARCHEDgoogle trends region

No stable official public API was used in the frozen research pass.

FETCH FAILEDgdelt failed request

The retained research expansion records one JSON-decode failure.

OUTSIDE THIS PAGEpartial 2026 attention

The captured 2026 year is partial and excluded from comparisons.

CLOSING READING

Privacy is not one thing moving in one direction.

The selected attention record and the institutional record meet at an operating surface: people encounter privacy as an idea, while documents and institutions translate it into notices, choices, duties, and risks.

SOURCES / RIGHTS

Attention figures use frozen Wikimedia Pageviews API responses for ten English Wikipedia pages with complete 2018–2025 coverage.

Policy language uses exact phrase counts from five frozen public policy or regulator documents that pass the 1,000-token capture floor. Upstream text retains its original rights.

Institutional anchors link to the named public sources. Research, data transforms, writing and design are by Dai Pan / 潘岱.

BACK TO TOP