Correlogram | Zyeon UI

Preview in your theme

status="ready" · glyph="ellipse" · layout="mixed" · defaultOrder="cluster" — three groups assemble on the diagonal, and the tilt of every ellipse is its sign.

Product metric correlationsShapes below the diagonal, numbers above it — the same 34 coefficients, twice9 variables · 34 of 36 pairs measured · strongest Sessions × Signups at 0.88 · 10 above p 0.05, kept and hatched

Order

Variable	TRL	SGN	SES	TKT	LAT	ACT	R30	CHN	NPS
Trial starts	TRL	0.84	0.79	0.19	0.07	0.52	0.38	-0.16	0.17
Signups		SGN	0.88	0.26	0.14	0.45	0.29	-0.11	0.09
Sessions			SES	0.31	0.22	0.41	0.33	-0.08	0.12
Support tickets				TKT	0.68	-0.44	-0.55	0.57
p95 latency					LAT	-0.51	-0.47	0.49
Activation						ACT	0.76	-0.69	0.63
Retention D30							R30	-0.82	0.71
Churn								CHN	-0.74
NPS									NPS

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

Pearson r over the trailing 90 days. The matrix is symmetric by construction: one coefficient per unordered pair, mirrored into both triangles. The lower triangle draws the shape, the other prints the number — same values, two encodings. Sign is geometry, not only hue: the ellipse leans / for positive and \ for negative. Rows and columns are in clustering order, so related variables sit next to each other and read as blocks on the diagonal.

Identical data, defaultOrder="input" — the order the query returned. Same coefficients, no visible structure: the reorder is the analysis, not a decoration.

Product metric correlations (as given)Press Cluster to watch acquisition, account health and friction separate9 variables · 34 of 36 pairs measured · strongest Sessions × Signups at 0.88 · 10 above p 0.05, kept and hatched

Order

Variable	SES	SGN	TRL	ACT	R30	NPS	TKT	LAT	CHN
Sessions	SES	0.88	0.79	0.41	0.33	0.12	0.31	0.22	-0.08
Signups		SGN	0.84	0.45	0.29	0.09	0.26	0.14	-0.11
Trial starts			TRL	0.52	0.38	0.17	0.19	0.07	-0.16
Activation				ACT	0.76	0.63	-0.44	-0.51	-0.69
Retention D30					R30	0.71	-0.55	-0.47	-0.82
NPS						NPS			-0.74
Support tickets							TKT	0.68	0.57
p95 latency								LAT	0.49
Churn									CHN

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

glyph="pie" · glyphSide="upper" · no p-values in the payload at all — nothing is masked, because "untested" must not be published as "passed".

Satisfaction item correlationsWedges sweep clockwise for positive; this study reported ρ and never ran the tests6 variables · 15 of 15 pairs measured · strongest Would recommend × Intends to renew at 0.81

Order

Variable	SPD	EAS	REC	RNW	VAL	SUP
Perceived speed	SPD
Ease of use	0.61	EAS
Would recommend	0.54	0.66	REC
Intends to renew	0.47	0.58	0.81	RNW
Value for money	0.21	0.29	0.57	0.62	VAL
Support quality	0.27	0.34	0.49	0.44	0.38	SUP

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

Spearman ρ on 5-point items. The matrix is symmetric by construction: one coefficient per unordered pair, mirrored into both triangles. The upper triangle draws the shape, the other prints the number — same values, two encodings. Sign is geometry, not only hue: the wedge sweeps clockwise for positive and anticlockwise for negative. Rows and columns are in clustering order, so related variables sit next to each other and read as blocks on the diagonal.

glyph="bar" · layout="glyphs" · cellSize={30} — 16 channels, 256 tiles. Past about a dozen variables the numbers stop fitting and the grid becomes a texture; a bar still resolves at 30px.

Channel cross-correlationOne pair in thirteen was never computed — hatched, not blank, and counted below16 variables · 111 of 120 pairs measured · strongest Channel 02 × Channel 03 at -0.91 · 45 above p 0.05, kept and hatched

Order

Variable	C10	C05	C09	C02	C03	C14	C04	C08	C15	C12	C01	C16	C06	C11	C13	C07
Channel 10	C10
Channel 05		C05
Channel 09			C09
Channel 02				C02
Channel 03					C03
Channel 14						C14
Channel 04							C04
Channel 08								C08
Channel 15									C15
Channel 12										C12
Channel 01											C01
Channel 16												C16
Channel 06													C06
Channel 11														C11
Channel 13															C13
Channel 07																C07

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

Pearson r over 30 days of one-minute samples. The matrix is symmetric by construction: one coefficient per unordered pair, mirrored into both triangles. Sign is geometry, not only hue: the bar runs right for positive and left for negative. Rows and columns are in clustering order, so related variables sit next to each other and read as blocks on the diagonal.

layout="numbers" · alpha={0.001} — a stricter threshold hatches more of the same grid. Masked cells keep their number, so the reader can still disagree with the threshold.

Product metric correlations (strict)Same coefficients as the first card, judged against p < 0.0019 variables · 34 of 36 pairs measured · strongest Sessions × Signups at 0.88 · 16 above p 0.001, kept and hatched

Order

Variable	TRL	SGN	SES	TKT	LAT	ACT	R30	CHN	NPS
Trial starts	TRL	0.84	0.79	0.19	0.07	0.52	0.38	-0.16	0.17
Signups	0.84	SGN	0.88	0.26	0.14	0.45	0.29	-0.11	0.09
Sessions	0.79	0.88	SES	0.31	0.22	0.41	0.33	-0.08	0.12
Support tickets	0.19	0.26	0.31	TKT	0.68	-0.44	-0.55	0.57
p95 latency	0.07	0.14	0.22	0.68	LAT	-0.51	-0.47	0.49
Activation	0.52	0.45	0.41	-0.44	-0.51	ACT	0.76	-0.69	0.63
Retention D30	0.38	0.29	0.33	-0.55	-0.47	0.76	R30	-0.82	0.71
Churn	-0.16	-0.11	-0.08	0.57	0.49	-0.69	-0.82	CHN	-0.74
NPS	0.17	0.09	0.12			0.63	0.71	-0.74	NPS

-1.00-0.60-0.250.250.601.00

p above 0.001 — outlined and hatched, never removednot computed

Pearson r over the trailing 90 days. The matrix is symmetric by construction: one coefficient per unordered pair, mirrored into both triangles. Sign is geometry, not only hue: the ellipse leans / for positive and \ for negative. Rows and columns are in clustering order, so related variables sit next to each other and read as blocks on the diagonal.

Degenerate payload: a repeated variable id, a NaN coefficient, one outside [-1, 1], an undeclared variable, a self-pair and a mirrored duplicate. One row of six survives — the rest are counted out loud.

One usable pair out of six rowsNothing is dropped quietly; every rejected row is named under the grid3 variables · 1 of 3 pairs measured · strongest Alpha × Beta at 0.62

Order

Variable	ALF	BET	GAM
Alpha	ALF	0.62
Beta		BET
Gamma			GAM

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

Pearson r. The matrix is symmetric by construction: one coefficient per unordered pair, mirrored into both triangles. The lower triangle draws the shape, the other prints the number — same values, two encodings. Sign is geometry, not only hue: the ellipse leans / for positive and \ for negative. Rows and columns are in clustering order, so related variables sit next to each other and read as blocks on the diagonal.

5 rows were not drawn: 2 carried a coefficient that is not a finite number inside [-1, 1]; 1 named a variable that was never declared; 1 joined a variable to itself; 1 repeated a pair that had already been given (the first one wins). Those cells read as not computed.

A single variable — a legal matrix with no pair in it. The diagonal renders, Cluster goes aria-disabled instead of vanishing, and nothing divides by zero.

A single variableClustering needs at least three variables and one measured pair1 variable · no pair to measure

Order

Variable	ARR
Annual recurring revenue	ARR

-1.00-0.60-0.250.250.601.00

p above 0.05 — outlined and hatched, never removednot computed

status="empty" — the variable list resolved, no coefficients came back.

Product metric correlationsOne variable set on both axes, mirrored by construction

Nothing to correlate yet

A correlogram needs a variable list and the coefficients between them — one row per unordered pair, not a full square and not the diagonal.

status="loading" — a triangular skeleton, and no sr-only text anywhere near a table.

Product metric correlationsOne variable set on both axes, mirrored by construction

status="error" with onRetry — retried 0 times

Product metric correlationsOne variable set on both axes, mirrored by construction

Couldn't load these coefficients

The data source didn't respond.

Loading preview…

This component is part of Pro

The install command, prompt and concept map are included with a Pro license — one purchase unlocks every Pro component, forever.

Unlock with Pro

TRL

SGN

SES

TKT

LAT

ACT

R30

CHN

NPS

Trial starts

TRL

0.84

0.79

0.19

0.07

0.52

0.38

-0.16

0.17

Signups

SGN

0.88

0.26

0.14

0.45

0.29

-0.11

0.09

Sessions

SES

0.31

0.22

0.41

0.33

-0.08

0.12

Support tickets

TKT

0.68

-0.44

-0.55

0.57

p95 latency

LAT

-0.51

-0.47

0.49

Activation

ACT

0.76

-0.69

0.63

Retention D30

R30

-0.82

0.71

Churn

CHN

-0.74

NPS

SES

SGN

TRL

ACT

R30

NPS

TKT

LAT

CHN

Sessions

SES

0.88

0.79

0.41

0.33

0.12

0.31

0.22

-0.08

Signups

SGN

0.84

0.45

0.29

0.09

0.26

0.14

-0.11

Trial starts

TRL

0.52

0.38

0.17

0.19

0.07

-0.16

Activation

ACT

0.76

0.63

-0.44

-0.51

-0.69

Retention D30

R30

0.71

-0.55

-0.47

-0.82

NPS

-0.74

Support tickets

TKT

0.68

0.57

p95 latency

LAT

0.49

Churn

CHN

SPD

EAS

REC

RNW

VAL

SUP

Perceived speed

SPD

Ease of use

0.61

EAS

Would recommend

0.54

0.66

REC

Intends to renew

0.47

0.58

0.81

RNW

Value for money

0.21

0.29

0.57

0.62

VAL

Support quality

0.27

0.34

0.49

0.44

0.38

SUP

C10

C05

C09

C02

C03

C14

C04

C08

C15

C12

C01

C16

C06

C11

C13

C07

Channel 10

C10

Channel 05

C05

Channel 09

C09

Channel 02

C02

Channel 03

C03

Channel 14

C14

Channel 04

C04

Channel 08

C08

Channel 15

C15

Channel 12

C12

Channel 01

C01

Channel 16

C16

Channel 06

C06

Channel 11

C11

Channel 13

C13

Channel 07

C07

TRL

SGN

SES

TKT

LAT

ACT

R30

CHN

NPS

Trial starts

TRL

0.84

0.79

0.19

0.07

0.52

0.38

-0.16

0.17

Signups

0.84

SGN

0.88

0.26

0.14

0.45

0.29

-0.11

0.09

Sessions

0.79

0.88

SES

0.31

0.22

0.41

0.33

-0.08

0.12

Support tickets

0.19

0.26

0.31

TKT

0.68

-0.44

-0.55

0.57

p95 latency

0.07

0.14

0.22

0.68

LAT

-0.51

-0.47

0.49

Activation

0.52

0.45

0.41

-0.44

-0.51

ACT

0.76

-0.69

0.63

Retention D30

0.38

0.29

0.33

-0.55

-0.47

0.76

R30

-0.82

0.71

Churn

-0.16

-0.11

-0.08

0.57

0.49

-0.69

-0.82

CHN

-0.74

NPS

0.17

0.09

0.12

0.63

0.71

-0.74

NPS

ALF

BET

GAM

Alpha

ALF

0.62

Beta

BET

Gamma

GAM

ARR

Annual recurring revenue

ARR