TOPICALAUTHORITY.ORG TAO / ROOT

Descriptive Analysis

TAO / ANALYSIS SYSTEM · STATE DESCRIPTIONANL / 02 OBSERVATION BOUNDED
ANL / 02 · DESCRIPTION LAYER

Descriptive Analysis

Describe the state before explaining it. Descriptive analysis reconstructs what was observed, in which population, at what time and under which measurement rules—without smuggling cause, mechanism or prediction into the result.

01 / OBJECTResolve the thing
02 / SCOPEDeclare the boundary
03 / UNITFix the denominator
04 / STATESummarize observations
05 / LOSSExpose missingness
06 / OUTPUTBound the description
01 OPERATIONAL DEFINITION

The observed state,made explicit.

Descriptive analysis is not a weak version of explanation. It is the controlled operation that determines exactly what the evidence says before any explanatory operation begins.

Descriptive analysis organizes, reduces and represents observed data inside a declared boundary so that the state, composition, distribution and variation of the object can be inspected without claiming why that state exists.

The output must preserve the observation unit, eligible population, measurement time, transformation rules, exclusions and missing records. A number without these controls is not a stable description; it is an unanchored value.

IT CAN STATEWhat exists and how it is distributed.
IT CAN COMPAREAligned states when the basis remains equivalent.
IT CANNOT ESTABLISHCause, mechanism, intent or inevitability.
SAFE OUTPUTA bounded account of the observed state.
02 QUESTION SPACE

Six questions.Six different summaries.

The analytical question determines the correct representation. Count, composition, distribution, location, recurrence and change are not interchangeable views of the same thing.

Q / COUNT

How many?

Returns a count of eligible observed units. The eligible population and duplicate rule must be declared.

OUTPUT / FREQUENCY
Q / RATE

How often?

Relates an event count to exposure, time or another valid denominator. A raw count cannot substitute for a rate.

OUTPUT / NORMALIZED FREQUENCY
Q / SHAPE

How distributed?

Shows center, spread, concentration, tails and outliers rather than collapsing the population into one average.

OUTPUT / DISTRIBUTION
Q / MIX

How composed?

Expresses categories as parts of an explicitly complete or qualified whole while retaining unknown categories.

OUTPUT / COMPOSITION
Q / POSITION

Where located?

Places units inside a spatial, architectural, semantic or market territory using a declared coordinate system.

OUTPUT / LOCATION
Q / TIME

What changed?

Describes a delta only when states share compatible units, definitions, collection rules and time boundaries.

OUTPUT / BOUNDED CHANGE
03 DENOMINATOR CONTROL

Every percentagehas a population.

The numerator describes the observed condition. The denominator defines the universe to which the statement applies. Change either and the meaning changes—even if the displayed percentage remains identical.

MEASUREMENT GRAMMAR

Count ≠ rate ≠ share.

A count reports magnitude. A rate normalizes events against exposure. A share reports composition within a whole. A ratio compares quantities that may not form a whole.

CONTROL: never write “61% of pages” when the population was actually “61% of eligible canonical HTML URLs observed successfully at collection time.”
NUMERATOR / CONDITION MET146 rival-only queries
÷
DENOMINATOR / ELIGIBLE SET250 normalized queries
58.4% observed share — valid only for this normalized query set, location, language, collection window and inclusion rule. It does not describe the entire market.
04 STATE RENDERER

One dataset.Four legitimate views.

Switch the analytical operation. The source values remain illustrative and fixed; only the representation and the claim it supports change.

DESCRIPTIVE RENDERER / ILLUSTRATIVE DATACOUNT VIEW
OBSERVATION / FREQUENCY

Observed query states

n = 250
RIVAL ONLY146
SHARED82
TARGET ONLY22
UNKNOWN0

SOURCE NOTE: values are an interface example used to demonstrate analytical representation; they are not live market measurements.

SUPPORTED STATEMENT

Magnitude

Inside the declared 250-query set, 146 observations are rival-only, 82 are shared and 22 are target-only.

DO NOT INFERThe rival caused the gap, the uncovered queries are valuable, or the same distribution exists outside this set.
05 MISSINGNESS MAP

Nothing is notone single state.

Zero, absent, unavailable, not applicable and excluded encode different realities. Collapsing them into one empty cell corrupts counts, shares and every later interpretation.

NULL-STATE TAXONOMY

Preserve why a value is missing.

A descriptive system must carry null-state provenance. If absence is transformed into zero, the interface creates evidence that was never observed.

0 / ZEROMeasured none

The property was eligible and measured; the resulting value is zero.

VALID NUMERIC STATE
∅ / ABSENTProperty not found

The expected property was not present within the inspected boundary.

OBSERVED ABSENCE
? / UNAVAILABLECould not observe

The value may exist, but collection failed or access was unavailable.

UNKNOWN STATE
— / N.A.Not applicable

The measure does not logically apply to this unit or category.

OUTSIDE MEASURE
× / EXCLUDEDRemoved by rule

The record existed but was excluded through a declared eligibility rule.

CONTROLLED REMOVAL
06 TEMPORAL DEPTH

A snapshot is nota trend.

Move the control from one to five observations. Descriptive language must change as temporal depth increases, while causal language remains prohibited.

OBSERVATION DEPTH CONTROL1 STATE
TEMPORAL CLAIM LIMIT

Snapshot

One bounded observation describes the state at collection time. It establishes neither direction nor persistence.

SUPPORTED LANGUAGE“At T1, the observed share was 58.4%.”
1 OBSERVATIONSnapshot
2 OBSERVATIONSDifference
3–5 OBSERVATIONSSequence / candidate trend
07 SUMMARY STATISTICS

The average cannotcarry the whole state.

Center, spread, range and composition expose different properties. A defensible description uses the smallest set of summaries that preserves the structure relevant to the question.

CENTER

Typical position

Mean, median and mode answer different questions and respond differently to skew and extreme values.

SPREAD

Variation

Range, interquartile range and standard deviation describe how far observations extend around the center.

POSITION

Quantiles

Percentiles locate an observation inside the distribution without pretending that intervals are equal in meaning.

COMPOSITION

Parts of the whole

Category shares require mutually intelligible groups, an explicit whole and preservation of unknown states.

08 WORKED DESCRIPTION

From records toa bounded statement.

Two domains, one grammar. The interface changes the object and values while preserving scope, unit, denominator, missingness and claim limits.

EXAMPLE / ASSET STATE

Describe the footprint.

The unit is one canonical HTML URL. The population contains every eligible URL discovered inside the declared domain snapshot.

BOUNDARYOne domain / public HTML / one crawl
UNITCanonical URL
DENOMINATOR240 eligible URLs
UNKNOWN22 unavailable responses retained
MEASUREVALUESHARE
Observed successfully21890.8%
Indexable18777.9%
Non-indexable3112.9%
Unavailable229.2%
BOUNDED DESCRIPTION: At collection time, 187 of 240 eligible URLs were observed as indexable. Twenty-two eligible URLs were unavailable and remain unknown; the result does not establish why any URL held its state.
09 FAILURE SURFACE

Descriptions failbefore explanations begin.

Most descriptive errors are boundary errors disguised as clean numbers. These eight failure modes contaminate every downstream comparison, pattern and conclusion.

ERR / 01

Percentage without denominator

The reported share cannot be mapped back to an eligible population.

ERR / 02

Mean without distribution

A center value hides skew, spread, clusters and extreme observations.

ERR / 03

Mixed categories

Categories overlap, change definition or combine unlike observation units.

ERR / 04

Snapshot as trend

One state or one difference is narrated as persistent directional movement.

ERR / 05

Rank as quality

An observed position is silently converted into an explanation or value judgment.

ERR / 06

Missing as zero

Unavailable evidence is recoded as a measured absence and changes the state.

ERR / 07

Aggregation blindness

A total conceals materially different segments, cohorts or territories.

ERR / 08

Stale description

A valid historical state is presented as if it still describes the current object.

10 EXECUTION PROTOCOL

Eight controls.One stable state.

The descriptive chain remains auditable because every reduction can be traced back to a declared object, source record and transformation.

01Resolve the object

Define identity and observation unit before collecting or grouping records.

CONTROL / IDENTITY
02Declare the boundary

Fix population, locale, surface, period, exclusions and eligibility rules.

CONTROL / SCOPE
03Preserve raw states

Retain source values, timestamps, provenance and original null conditions.

CONTROL / SOURCE
04Normalize explicitly

Apply documented mappings without erasing materially different states.

CONTROL / TRANSFORM
05Fix denominators

Bind each count, rate and share to its valid eligible population.

CONTROL / UNIT
06Summarize structure

Select measures that preserve center, spread, composition or location.

CONTROL / REDUCTION
07Audit missingness

Separate zero, absence, unavailable, not applicable and excluded records.

CONTROL / LOSS
08Bound the statement

State what was observed, where and when—and what was not established.

CONTROL / CLAIM
11 DECISION GATE

Complete description.No causal leakage.

A descriptive output is ready for exploratory or comparative analysis only when another analyst can reconstruct the population, operations and claim boundary.

DESCRIPTIVE OUTPUT CONTRACT
Within [boundary], at [time], [value or distribution] was observed across [eligible population], subject to [missingness and exclusions].

This grammar keeps the statement attached to its evidence. Explanation, mechanism, forecast and decision value remain separate analytical operations.

Object identity and observation unit declared
Population, scope and collection time explicit
Numerator and denominator reconstructable
Missing, excluded and unavailable records visible
Transformations and category rules traceable
No causal, predictive or universal claim introduced
12 ANALYSIS ROUTER

One root.Twelve analytical nodes.

Descriptive Analysis is the second node: it converts governed observations into a bounded account of state before discovery, comparison, pattern detection or causal reasoning.

TOPICALAUTHORITY.ORG / ANALYSIS SYSTEMANL / 02 · DESCRIPTION BOUNDED
TAO / CONTACT · DIRECT TRANSMISSION Have an asset, domain or market position to investigate? ENTER CONTACT SYSTEM →
DIGITAL ASSET INTELLIGENCE + EXECUTION
EXECUTED BY
BB DIGITALNA AGENCIJA

Investigation, consulting and execution of digital assets, premium-domain strategies, information architecture, semantic systems, websites and agreed digital growth plans.

TOPICALAUTHORITY.ORG / SEMANTIC INTELLIGENCE SYSTEM BB DIGITALNA AGENCIJA / BB.HR