Exact relationship builder
Indicator comparison lab
Compare two measures while keeping their entities, periods, units, and source populations visible. A visual association does not establish causation.
Market structure triangulation
Read concentration through three bounded measures
Switch among every exact pair without treating HHI, CR4, and leader share as interchangeable or additive.52 exact pair rowsThree declared edges / four released periods
Each edge is an independent descriptive contract. Shared arithmetic is disclosed, missing sides stay missing, and no three-measure composite score is produced.
Paired visualization
Tracked market CR4 and leading-participant share
4 exact pairs in this slice- Exact pairs
- 4
- all plotted
- Correlation
- +0.68
- Moderate positive association
- R²
- 0.46
- descriptive linear fit
- Evidence
- 0
- verified on both sides
Paired comparison cohort
0 / 5 entities · fixed 4-pair benchmarkSelect a point, then add it or its nearest exact-pair peers.
The cohort changes the overlay and exact export only. Full-slice medians, correlation, quadrants, and coverage remain fixed.
- Market CR4
- 99.9%
- Estimated public · Medium
- Market leader share
- 56.5%
- Estimated public · Medium
- X source
- statcounter-mobile-os-us-2026
- Y source
- statcounter-mobile-os-us-2026
Coverage diagnostics
What entered the plane, and what did not
Pairing coverage is reported separately for each indicator denominator. A matched row is not evidence that either source covers the full underlying economy.
Only common-release observations are paired. Leader share is a component of CR4, so association is partly mechanical; the plane describes how top-four concentration is distributed and does not independently establish market power or a legal market finding.
Quadrants use this filtered slice's medians: 97.5% and 54.7%.
Slice ledgers
Period and group structure
Counts and correlations stay inside exact paired rows. Small groups may be too thin for a stable descriptive association.
| Group | Pairs | X avg. | Y avg. | r |
|---|---|---|---|---|
| Digital platforms | 3 | 96.8% | 63.0% | +0.67 |
| Software | 1 | 95.5% | 53.0% | N/A |
Reusable comparison packet
Carry the exact contract into analysis
The endpoint returns every matched row in the filtered slice, both observation identifiers, both source routes, evidence status, confidence, entity, group, and period.
This lab compares only declared same-entity, same-period pairs. Missing sides are excluded rather than imputed, correlation remains descriptive, and unlike universes are never combined into a composite market-power score.
Relationship directory
15 declared planes, no arbitrary joins
Each route names both measures and its exact pairing rule. Select a plane to rebuild the lab around that bounded universe.