Chart Scatter Matrix

status="ready" · 5 variables = 25 panels: scatter below the diagonal, distribution histograms on it, Pearson r above. Bring the pointer near any point (nearest one within 10 units) and that record lights up in every panel at once; Tab into the grid and the arrow keys walk record by record.

Service fleet · five signals crossed

54 services · last 24h · three regions

Strongest linear relationship: p95 latency vs Requests, r = +0.97 on 54 complete pairs. 3 records missing at least one value — left out of the panels that need it, never plotted as 0. r is Pearson's, on complete pairs only: it measures straight-line association and nothing else — all four of Anscombe's very different clouds score 0.82. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Service fleet · five signals crossed: 5 variables — p95 latency, Error rate, CPU, Requests, Cost per 1k — crossed against each other over 54 records, 10 scatter panels with each variable's distribution on the diagonal. Strongest linear relationship: p95 latency and Requests, r = +0.97 over 54 complete pairs. 3 records are missing at least one value and sit out of the panels that need it. Every value follows in the data table.
Record	Series	p95 latency (ms)	Error rate (%)	CPU (%)	Requests (/min)	Cost per 1k (USD)
checkout-a	us-east	240 ms	0%	61%	2,811/min	0.77 USD
checkout-b	us-east	314 ms	0.05%	66%	3,288/min	0.619 USD
checkout-c	us-east	401 ms	0.89%	84%	3,935/min	0.493 USD
search-a	us-east	379 ms	0.77%	73%	3,713/min	0.592 USD
search-b	us-east	425 ms	1.14%	81%	3,993/min	0.519 USD
search-c	us-east	462 ms	1.42%	93%	4,287/min	0.603 USD
auth-a	us-east	444 ms	1.29%	83%	4,194/min	0.632 USD
auth-b	us-east	60 ms	0.01%	33%	1,413/min	1.143 USD
auth-c	us-east	108 ms	0.01%	35%	1,347/min	1.179 USD
catalog-a	us-east	87 ms	0.02%	30%	1,285/min	1.123 USD
catalog-b	us-east	370 ms	0.74%	73%	4,052/min	0.513 USD
catalog-c	us-east	279 ms	0%	64%	3,446/min	0.747 USD
cart-a	us-east	107 ms	0%	47%	1,680/min	1.159 USD
cart-b	us-east	182 ms	0%	61%	2,812/min	not measured
cart-c	us-east	224 ms	0.12%	56%	3,106/min	0.818 USD
payments-a	us-east	209 ms	0.06%	56%	2,900/min	0.915 USD
payments-b	us-east	294 ms	0.09%	66%	3,260/min	0.602 USD
payments-c	us-east	329 ms	0.16%	66%	3,470/min	0.72 USD
shipping-a	us-east	98 ms	0.02%	47%	1,788/min	1.038 USD
shipping-b	us-east	76 ms	0%	23%	905/min	1.2 USD
shipping-c	us-east	276 ms	0.25%	67%	3,462/min	0.677 USD
inventory-a	us-east	386 ms	0.82%	85%	3,692/min	0.526 USD
inventory-b	eu-west	104 ms	0.03%	46%	1,893/min	1.053 USD
inventory-c	eu-west	357 ms	0.47%	72%	3,524/min	0.622 USD
recommend-a	eu-west	77 ms	0.07%	28%	1,037/min	1.105 USD
recommend-b	eu-west	331 ms	0.48%	81%	3,633/min	0.714 USD
recommend-c	eu-west	72 ms	0.03%	35%	1,499/min	1.229 USD
reviews-a	eu-west	464 ms	1.27%	91%	4,339/min	0.586 USD
reviews-b	eu-west	310 ms	0.3%	75%	3,521/min	0.663 USD
reviews-c	eu-west	449 ms	1.4%	88%	4,252/min	0.503 USD
media-a	eu-west	197 ms	0.02%	64%	2,782/min	not measured
media-b	eu-west	398 ms	1.01%	76%	3,804/min	0.675 USD
media-c	eu-west	192 ms	0.07%	56%	2,605/min	0.86 USD
notify-a	eu-west	76 ms	0.08%	32%	1,559/min	1.006 USD
notify-b	eu-west	449 ms	1.36%	85%	4,123/min	0.546 USD
notify-c	eu-west	357 ms	0.69%	82%	3,765/min	0.622 USD
billing-a	eu-west	86 ms	0.02%	44%	1,493/min	1.072 USD
billing-b	eu-west	496 ms	1.6%	88%	4,351/min	0.457 USD
billing-c	eu-west	402 ms	0.85%	76%	3,714/min	0.531 USD
identity-a	eu-west	208 ms	0%	61%	2,732/min	0.961 USD
identity-b	ap-south	244 ms	0.14%	68%	2,819/min	0.82 USD
identity-c	ap-south	263 ms	0.08%	61%	3,015/min	0.892 USD
pricing-a	ap-south	34 ms	0.07%	33%	765/min	1.171 USD
pricing-b	ap-south	397 ms	0.76%	78%	3,651/min	0.641 USD
pricing-c	ap-south	200 ms	0.09%	57%	2,823/min	0.933 USD
tracking-a	ap-south	291 ms	0.26%	71%	3,314/min	0.74 USD
tracking-b	ap-south	451 ms	1.13%	84%	4,168/min	0.616 USD
tracking-c	ap-south	117 ms	0.14%	48%	1,954/min	1.1 USD
support-a	ap-south	336 ms	0.37%	66%	3,511/min	0.638 USD
support-b	ap-south	344 ms	0.59%	77%	3,872/min	0.598 USD
support-c	ap-south	78 ms	0.03%	38%	1,040/min	not measured
reports-a	ap-south	247 ms	0.02%	57%	2,616/min	0.771 USD
reports-b	ap-south	31 ms	0.05%	33%	782/min	1.264 USD
reports-c	ap-south	377 ms	0.9%	80%	3,774/min	0.601 USD

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
p95 latency vs Error rate	+0.85	54
p95 latency vs CPU	+0.96	54
p95 latency vs Requests	+0.97	54
p95 latency vs Cost per 1k	-0.95	51
Error rate vs CPU	+0.80	54
Error rate vs Requests	+0.76	54
Error rate vs Cost per 1k	-0.74	51
CPU vs Requests	+0.97	54
CPU vs Cost per 1k	-0.93	51
Requests vs Cost per 1k	-0.96	51

status="empty" · a valid payload with not one row in it — kept apart from "still loading" and "the fetch failed".

Service fleet · five signals crossed

54 services · last 24h · three regions

Nothing to cross-plot yet

A scatter matrix needs at least one variable and one record carrying a measured value.

status="loading" · a 4×4 skeleton of panels, aria-hidden, with no timer behind it.

Service fleet · five signals crossed

54 services · last 24h · three regions

status="error" · the fetch failed; Try again shows up only when onRetry is passed, and pressing it switches to ready.

Service fleet · five signals crossed

54 services · last 24h · three regions

Couldn't load the matrix

The data source didn't respond.

Degenerate data · Plan tier is 3 on every row (the domain is invented, so the panel reads “no spread” and r for that column is simply not defined); Margin straddles 0, so the panel draws a dashed zero axis; ARR runs into the billions behind a label far too long for its panel, so its ticks go compact (“-30.7M” is wide enough to collide with the middle one, so that column drops its middle tick and always keeps both ends) and the title truncates on real glyph metrics inside a foreignObject; two rows have holes — a missing cell is left undrawn, never drawn as 0.

Accounts · awkward columns

One flat column, one that crosses zero, one in the billions, two holes

Strongest linear relationship: Margin vs Annual recurring revenue before discounts, reporting currency, r = +0.52 on 6 complete pairs. 1 record missing at least one value — left out of the panels that need it, never plotted as 0. r is Pearson's, on complete pairs only: it measures straight-line association and nothing else — all four of Anscombe's very different clouds score 0.82. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Accounts · awkward columns: 4 variables — Margin, Plan tier, Annual recurring revenue before discounts, reporting currency, Seats — crossed against each other over 7 records, 6 scatter panels with each variable's distribution on the diagonal. Strongest linear relationship: Margin and Annual recurring revenue before discounts, reporting currency, r = +0.52 over 6 complete pairs. 1 record is missing at least one value and sits out of the panels that need it. Every value follows in the data table.
Record	Margin (%)	Plan tier	Annual recurring revenue before discounts, reporting currency (USD)	Seats
Northwind	-1.8%	3	1,240,000,000 USD	120
Contoso	4.2%	3	880,500,000 USD	340
Fabrikam	0%	3	2,410,000,000 USD	90
Tailspin	-6.4%	3	145,900,000 USD	610
Proseware	9.1%	3	not measured	not measured
Litware	2.7%	3	3,090,000,000 USD	780
Adventure Works	-0.4%	3	512,300,000 USD	260

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
Margin vs Plan tier	not defined	7
Margin vs Annual recurring revenue before discounts, reporting currency	+0.52	6
Margin vs Seats	-0.01	6
Plan tier vs Annual recurring revenue before discounts, reporting currency	not defined	6
Plan tier vs Seats	not defined	6
Annual recurring revenue before discounts, reporting currency vs Seats	+0.17	6

The three sets of Anscombe's quartet that share an x column: every pair scores r = +0.82 against x, and no two scatters are the same picture (a line, a curve, one point holding it up). This is the original proof that a correlation coefficient on its own will lie to you — and the reason a scatter matrix exists.

Anscombe's quartet

Every pair with x scores r = +0.82 — and none of them is the same shape

Strongest linear relationship: x (shared) vs Set 1 · linear, r = +0.82 on 11 complete pairs. r is Pearson's, on complete pairs only: it measures straight-line association and nothing else — all four of Anscombe's very different clouds score 0.82. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Anscombe's quartet: 4 variables — x (shared), Set 1 · linear, Set 2 · curved, Set 3 · one point — crossed against each other over 11 records, 6 scatter panels with each variable's distribution on the diagonal. Strongest linear relationship: x (shared) and Set 1 · linear, r = +0.82 over 11 complete pairs. Every value follows in the data table.
Record	x (shared)	Set 1 · linear	Set 2 · curved	Set 3 · one point
Observation 1	10	8.04	9.14	7.46
Observation 2	8	6.95	8.14	6.77
Observation 3	13	7.58	8.74	12.74
Observation 4	9	8.81	8.77	7.11
Observation 5	11	8.33	9.26	7.81
Observation 6	14	9.96	8.1	8.84
Observation 7	6	7.24	6.13	6.08
Observation 8	4	4.26	3.1	5.39
Observation 9	12	10.84	9.13	8.15
Observation 10	7	4.82	7.26	6.42
Observation 11	5	5.68	4.74	5.73

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
x (shared) vs Set 1 · linear	+0.82	11
x (shared) vs Set 2 · curved	+0.82	11
x (shared) vs Set 3 · one point	+0.82	11
Set 1 · linear vs Set 2 · curved	+0.75	11
Set 1 · linear vs Set 3 · one point	+0.47	11
Set 2 · curved vs Set 3 · one point	+0.59	11

One record only · a single mark per panel; with fewer than 3 complete pairs Pearson r is reported as not defined instead of a confident ±1 (any two points are collinear by construction).

Day one

One record, three variables — nothing to correlate, still worth drawing

r is Pearson's, on complete pairs only: it measures straight-line association and nothing else — all four of Anscombe's very different clouds score 0.82. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Day one: 3 variables — p95 latency, Error rate, Requests — crossed against each other over 1 record, 3 scatter panels with each variable's distribution on the diagonal. No pair has enough complete, varying data for a correlation. Every value follows in the data table.
Record	p95 latency (ms)	Error rate (%)	Requests (/min)
First measurement	214 ms	0.4%	1,180/min

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
p95 latency vs Error rate	not defined	1
p95 latency vs Requests	not defined	1
Error rate vs Requests	not defined	1

status="ready" with points empty — the columns are known, the rows have not landed. The component falls back to its empty branch instead of drawing a blank grid that pretends to hold data.

Fresh workspace

Columns are known, the first row has not landed

Nothing to cross-plot yet

A scatter matrix needs at least one variable and one record carrying a measured value.

Fed 12 variables × 900 rows (144 panels, roughly 59,000 marks in the scatter halves alone) · maxVariables caps at 5 and maxPoints at 250, and the card footer states what was drawn and what was left out rather than grinding to a halt drawing all of it.

Feature sweep

Twelve variables and nine hundred rows offered; the caps decide what is drawn

Strongest linear relationship: Feature 01 vs Feature 04, r = +0.83 on 250 complete pairs. Drawing 5 of 12 variables — 25 panels. Not drawn: Feature 06, Feature 07, Feature 08, Feature 09, Feature 10, Feature 11, Feature 12. Drawing the first 250 of 900 records. r is Pearson's, on complete pairs only: it measures straight-line association and nothing else — all four of Anscombe's very different clouds score 0.82. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Feature sweep: 5 variables — Feature 01, Feature 02, Feature 03, Feature 04, Feature 05 — crossed against each other over 250 records, 10 scatter panels with each variable's distribution on the diagonal. Strongest linear relationship: Feature 01 and Feature 04, r = +0.83 over 250 complete pairs. 7 further variables not drawn: Feature 06, Feature 07, Feature 08, Feature 09, Feature 10, Feature 11, Feature 12. 650 further records not drawn. Every value follows in the data table.
Record	Series	Feature 01	Feature 02	Feature 03	Feature 04	Feature 05
Row 1	Cohort 1	79	43	60.8	83.2	40.3
Row 2	Cohort 2	51.8	43.9	53.6	56.8	41
Row 3	Cohort 3	90	25.1	59.4	88.4	23
Row 4	Cohort 4	59.3	54.4	42.5	53.1	54.8
Row 5	Cohort 1	79.6	36.1	55.6	82.7	28.8
Row 6	Cohort 2	50	49.8	58.9	49.5	44.7
Row 7	Cohort 3	84.9	32.7	61.8	84.5	35.5
Row 8	Cohort 4	81.4	34.2	53.8	79	26.7
Row 9	Cohort 1	76.6	42.7	58.7	66.7	34.4
Row 10	Cohort 2	82.5	33.4	60.2	78.4	30.8
Row 11	Cohort 3	54.8	47.4	55.5	67.7	44.4
Row 12	Cohort 4	83	30.1	50.1	73	28.5
Row 13	Cohort 1	71.2	35.5	51.4	66.8	31.5
Row 14	Cohort 2	48.7	43.1	51.3	63.7	54.8
Row 15	Cohort 3	53.1	36.5	50.2	51.3	36.7
Row 16	Cohort 4	66.4	35.3	54.6	67.3	44.4
Row 17	Cohort 1	79.5	21.2	46.5	84.2	33.4
Row 18	Cohort 2	50.8	39.6	52.8	46.9	50.4
Row 19	Cohort 3	63.6	35.7	50.5	61.2	40.4
Row 20	Cohort 4	67.2	48.3	57.5	68.3	45.2
Row 21	Cohort 1	78.6	39.4	62.1	88.6	26.2
Row 22	Cohort 2	80.9	29.5	44.6	68.5	37.4
Row 23	Cohort 3	75.9	34.8	48.9	89.5	34.2
Row 24	Cohort 4	82.9	29.7	47.4	87.9	34.5
Row 25	Cohort 1	81.5	33.1	61.7	77.5	38.7
Row 26	Cohort 2	48.9	49.7	53.7	55.4	48.9
Row 27	Cohort 3	86.1	26.9	58.5	80.6	37.9
Row 28	Cohort 4	79.3	26.6	56.8	75.3	42.1
Row 29	Cohort 1	53.2	49.3	52.5	50.2	52.5
Row 30	Cohort 2	85.6	24.6	52.2	83.8	19.6
Row 31	Cohort 3	67.9	41.6	61	70.8	41.4
Row 32	Cohort 4	83.8	21.9	55.5	86	33.6
Row 33	Cohort 1	75.1	25.6	53.4	78	26.4
Row 34	Cohort 2	77	27.1	47.9	74.9	41.6
Row 35	Cohort 3	47.9	52.2	45.3	50.4	45.2
Row 36	Cohort 4	63.1	47.3	45.5	55.7	49.6
Row 37	Cohort 1	57.5	34.9	52.6	62.7	46.8
Row 38	Cohort 2	68	36	49.3	57.7	35.5
Row 39	Cohort 3	51.8	55.2	41.9	57.6	57.4
Row 40	Cohort 4	69.1	36.1	56.8	64.1	42.8
Row 41	Cohort 1	64.8	41.6	57.8	70.3	37.2
Row 42	Cohort 2	57.7	49.4	43.8	61.3	44.5
Row 43	Cohort 3	94.9	30.5	56.7	91.9	23.5
Row 44	Cohort 4	83.1	23.2	62.5	93.1	30
Row 45	Cohort 1	80.6	23.1	48.8	81.1	22.2
Row 46	Cohort 2	64.9	46	53.7	62.3	47.4
Row 47	Cohort 3	72.7	39.6	56.7	85.4	32.7
Row 48	Cohort 4	52.3	37.2	54.3	63	43.7
Row 49	Cohort 1	53.4	41.8	52.7	53.8	42.2
Row 50	Cohort 2	58.6	40.4	48.9	60.8	39.8
Row 51	Cohort 3	62.3	38.5	54.6	77.2	32.6
Row 52	Cohort 4	68.2	50.1	55.7	60.3	52
Row 53	Cohort 1	86.8	33	53.4	84.1	19.1
Row 54	Cohort 2	76.7	41.9	57.8	77.6	47.4
Row 55	Cohort 3	59	47.4	41.3	47.5	44.8
Row 56	Cohort 4	80.8	31.9	50.6	80.4	25.9
Row 57	Cohort 1	66.4	33.3	52.4	65	28.3
Row 58	Cohort 2	63.4	44.3	61.6	78	35.4
Row 59	Cohort 3	74.8	31.8	44.2	76.8	30.4
Row 60	Cohort 4	69.2	30.8	59.3	77.9	41.1
Row 61	Cohort 1	77.4	29.4	48.6	78.6	33.8
Row 62	Cohort 2	77	30.7	52.5	65.5	32
Row 63	Cohort 3	94.4	28.9	52.5	82.5	20.7
Row 64	Cohort 4	72.5	33.3	58.6	68.2	35.7
Row 65	Cohort 1	70.8	37.9	58.9	66.1	47.5
Row 66	Cohort 2	76.6	31.7	58.8	82	39
Row 67	Cohort 3	90.1	34	46.1	87.7	36.7
Row 68	Cohort 4	50.8	52.3	46.7	56.5	55.5
Row 69	Cohort 1	88.4	36.6	53.1	80.3	28.7
Row 70	Cohort 2	64.7	49.1	60.8	67.3	47.7
Row 71	Cohort 3	74.1	34.1	60.8	80.2	27.6
Row 72	Cohort 4	74.6	23.1	53.3	75.8	32.5
Row 73	Cohort 1	81.7	26.8	57.2	71.7	24.9
Row 74	Cohort 2	66.2	31	55.7	77.6	36.2
Row 75	Cohort 3	66.2	47.1	47	60.2	50.2
Row 76	Cohort 4	65.1	33.1	59.2	75.1	31.7
Row 77	Cohort 1	75.4	43.9	55.3	78.2	42.5
Row 78	Cohort 2	46.3	42.5	46.7	47.9	53
Row 79	Cohort 3	49.1	51.7	59.2	51.7	44.8
Row 80	Cohort 4	96.9	26	56.2	93.5	26.3
Row 81	Cohort 1	81	44	58.9	69	34.9
Row 82	Cohort 2	64	38.8	52.7	61	40
Row 83	Cohort 3	84.8	25.4	59.1	84.2	29.6
Row 84	Cohort 4	61.3	43.1	58.8	59.2	50.7
Row 85	Cohort 1	60.4	49.4	47.7	65.1	34.7
Row 86	Cohort 2	69.6	40.4	59.1	83.3	34.8
Row 87	Cohort 3	50.1	44.8	46.1	46.9	57.5
Row 88	Cohort 4	76.8	31.1	47.3	75	34.5
Row 89	Cohort 1	76.3	28.4	51.7	79.3	34.5
Row 90	Cohort 2	73.3	39.8	61.5	69.1	36.8
Row 91	Cohort 3	58.2	42.5	44.5	46.9	49.4
Row 92	Cohort 4	48.4	46.2	49.8	52	44.9
Row 93	Cohort 1	61.6	48.3	59.4	50.1	45.6
Row 94	Cohort 2	78.5	29.1	46.1	72.5	44.3
Row 95	Cohort 3	68.5	41	59.4	70.6	34
Row 96	Cohort 4	78.5	38.9	50.5	86.1	26.7
Row 97	Cohort 1	85.6	27.3	46.7	88.4	32.7
Row 98	Cohort 2	67.8	41.8	58.8	74	35
Row 99	Cohort 3	86.4	23.3	50.3	78.4	35.3
Row 100	Cohort 4	73.3	36.6	60.4	78.6	35.8
Row 101	Cohort 1	68.8	34.6	59	71.5	43.3
Row 102	Cohort 2	68.5	34.5	55.9	77.8	42.9
Row 103	Cohort 3	61	32.7	43.6	71.3	36.9
Row 104	Cohort 4	75.8	38.5	46	68.6	34.6
Row 105	Cohort 1	80	36.6	49.2	76.9	25.3
Row 106	Cohort 2	50.2	38.2	57.9	58.6	39.9
Row 107	Cohort 3	55.5	54	51.7	53	39.6
Row 108	Cohort 4	69.6	39.2	47.2	68.4	38.6
Row 109	Cohort 1	72.4	40.1	59.2	66.1	44.5
Row 110	Cohort 2	74.5	30.9	60.3	85.6	25.7
Row 111	Cohort 3	77.4	20.7	59.1	78.8	24.8
Row 112	Cohort 4	67.8	42.8	53.2	58.3	39.1
Row 113	Cohort 1	84.1	35.7	48.7	81	31.5
Row 114	Cohort 2	52.4	45.5	53.7	59.6	55.7
Row 115	Cohort 3	81.4	25.6	53.4	80.7	33.8
Row 116	Cohort 4	96.6	33.5	59.4	83.5	25
Row 117	Cohort 1	73.9	32.2	47.9	77.9	26.8
Row 118	Cohort 2	54.7	49.5	55.8	51.4	48.6
Row 119	Cohort 3	63.5	47.4	52.9	58.3	50.4
Row 120	Cohort 4	57	38.6	55.8	58.8	47.2
Row 121	Cohort 1	62.5	52.3	54	51.4	43.8
Row 122	Cohort 2	61.6	36.2	58.9	64.8	39.2
Row 123	Cohort 3	54.4	48.3	56.8	52.3	43.1
Row 124	Cohort 4	53.9	50.4	53.8	57.4	44.1
Row 125	Cohort 1	88.5	19.3	55.8	88.1	36
Row 126	Cohort 2	54.2	39.1	59.1	54.5	44.8
Row 127	Cohort 3	56.6	49.9	56.3	52.1	51
Row 128	Cohort 4	78	31.1	60.4	67	39.2
Row 129	Cohort 1	73.1	42.7	62.4	80.6	26.2
Row 130	Cohort 2	56.5	39.5	50	70.1	45.3
Row 131	Cohort 3	75.9	31.3	45.4	65	30.5
Row 132	Cohort 4	55.8	45.1	53.3	71.1	40.2
Row 133	Cohort 1	65.1	46.7	47.3	69.9	30.9
Row 134	Cohort 2	66.4	46.2	60.6	69.3	43.8
Row 135	Cohort 3	66.4	43.5	44.1	56.5	51.4
Row 136	Cohort 4	83.1	38.3	52.8	78.5	27.1
Row 137	Cohort 1	61.1	43.9	48.6	58.1	56.3
Row 138	Cohort 2	76.8	24.3	46.7	83.1	29.7
Row 139	Cohort 3	83.3	27.3	56.2	84.6	26.4
Row 140	Cohort 4	72.1	45	60	63.5	42.7
Row 141	Cohort 1	70.1	37.7	56.5	77.7	31.8
Row 142	Cohort 2	76.3	38.4	45.2	76.9	38.3
Row 143	Cohort 3	68.2	41.1	46.1	73.1	39.1
Row 144	Cohort 4	48.6	53.5	58.4	53.1	55.7
Row 145	Cohort 1	54.5	46.6	42.2	53.5	51.8
Row 146	Cohort 2	67.1	49.2	49.3	69.5	40
Row 147	Cohort 3	62.7	53.5	42.1	62.4	52.5
Row 148	Cohort 4	71.1	38.7	60.4	80.1	24.4
Row 149	Cohort 1	84	37	54.8	82.3	28.8
Row 150	Cohort 2	79.9	40.6	57.5	80.2	36.6
Row 151	Cohort 3	88.6	36.5	54.5	81	28
Row 152	Cohort 4	57.8	41.5	60.3	61.3	37.1
Row 153	Cohort 1	55.9	49.9	55.6	45.4	41.4
Row 154	Cohort 2	46.1	49.8	42.6	45.6	51.6
Row 155	Cohort 3	77.2	41.4	53.9	70.8	28.5
Row 156	Cohort 4	74	37.2	44.3	69.1	39.4
Row 157	Cohort 1	74.5	39.7	46.9	65.3	36.6
Row 158	Cohort 2	78.3	29.5	52.6	85.2	39.4
Row 159	Cohort 3	66.9	40.2	54.8	77.9	26.3
Row 160	Cohort 4	63.1	43.7	48.4	58.4	38.3
Row 161	Cohort 1	94.4	20.1	60.8	85.8	22.5
Row 162	Cohort 2	92.4	33.1	52	84.3	19.2
Row 163	Cohort 3	70.6	45.5	48.5	74.1	42.4
Row 164	Cohort 4	73.2	44.2	45.3	74.8	38
Row 165	Cohort 1	65	37.2	47.4	75	32.7
Row 166	Cohort 2	86.2	36.7	49.9	75.6	37.7
Row 167	Cohort 3	65.5	44.2	58.8	64.3	37.8
Row 168	Cohort 4	83.2	34.4	52.6	75.4	28.4
Row 169	Cohort 1	78.1	37.1	56.3	83.9	39.3
Row 170	Cohort 2	89	32.2	58.1	79.7	38
Row 171	Cohort 3	80.1	37.6	55.5	82	34.1
Row 172	Cohort 4	55.9	34.3	50.4	68	35.2
Row 173	Cohort 1	70.4	30.9	61.9	71	38.8
Row 174	Cohort 2	83.2	38.4	48.7	71.5	37.1
Row 175	Cohort 3	73.2	35.7	55.3	62.3	47
Row 176	Cohort 4	91.4	29.9	56.4	84.1	38.4
Row 177	Cohort 1	58.2	34.2	52	75.2	33.5
Row 178	Cohort 2	87.1	22.3	60.2	84.1	19.6
Row 179	Cohort 3	93.6	20.2	55.1	94.6	32.3
Row 180	Cohort 4	67.3	34.3	61.3	62.5	33.3
Row 181	Cohort 1	86.5	25.5	49.6	77	34.4
Row 182	Cohort 2	79.9	27.3	46	69.7	28.4
Row 183	Cohort 3	61.5	45.3	52.1	63.3	38
Row 184	Cohort 4	50.1	38.6	53.5	63.1	47.3
Row 185	Cohort 1	58.9	52.4	42.3	52.8	54.8
Row 186	Cohort 2	81.3	26.4	62.1	75.8	30.3
Row 187	Cohort 3	47	44.1	44	50.5	53.2
Row 188	Cohort 4	73.9	32.3	46.2	72	45.8
Row 189	Cohort 1	81.4	39	46.5	73.1	22.6
Row 190	Cohort 2	70	48.2	49	59.1	38.2
Row 191	Cohort 3	69.4	42.4	48.5	61.9	37.5
Row 192	Cohort 4	85.3	28.3	49.8	87.5	31.3
Row 193	Cohort 1	83.3	35.7	54.9	75.9	34.7
Row 194	Cohort 2	71.5	28.5	54.1	70.4	40.9
Row 195	Cohort 3	75.7	34.6	62.4	68.5	28.4
Row 196	Cohort 4	59.4	39.2	60.4	59.8	37.6
Row 197	Cohort 1	86.3	32.2	55.7	86.7	34
Row 198	Cohort 2	70.6	30.6	52.1	64.3	33.9
Row 199	Cohort 3	92.2	31	60.3	76.2	23.5
Row 200	Cohort 4	69.1	42.5	51.9	63.1	37.8
Row 201	Cohort 1	50.4	39.4	53.3	49	41.9
Row 202	Cohort 2	58.7	39.9	56.5	76.2	34.8
Row 203	Cohort 3	82.5	31.3	62.6	86.1	32.8
Row 204	Cohort 4	66.2	38.3	60.4	75.5	42.9
Row 205	Cohort 1	61.7	42.4	59.8	73.8	37.1
Row 206	Cohort 2	94.7	25.2	63.7	92	19.7
Row 207	Cohort 3	70	40.1	56.3	78.8	43.3
Row 208	Cohort 4	78.6	46.5	56.8	68.2	41.5
Row 209	Cohort 1	59.4	37.1	49.6	69.1	33.5
Row 210	Cohort 2	75.9	27	46.1	75.1	39.9
Row 211	Cohort 3	62.4	33.4	49.4	72.7	41.3
Row 212	Cohort 4	70.1	38.3	51.3	62.9	39.6
Row 213	Cohort 1	46	52.8	48.2	45.2	50.8
Row 214	Cohort 2	73.7	31.5	45.8	76.3	44.1
Row 215	Cohort 3	85.2	33.6	51.3	82.1	36.6
Row 216	Cohort 4	70.9	38.4	59.3	67	27.7
Row 217	Cohort 1	81	25.8	45.2	68.1	32.5
Row 218	Cohort 2	61.2	42.8	60.7	74.6	47.3
Row 219	Cohort 3	92.3	29.3	61.2	76.3	35.4
Row 220	Cohort 4	50.4	36.7	42.7	55.5	37.3
Row 221	Cohort 1	52.9	37.3	49.9	53.5	36.4
Row 222	Cohort 2	46.4	50.2	41.9	52.4	45.1
Row 223	Cohort 3	88.1	34.8	57.2	79.4	28.1
Row 224	Cohort 4	88	23	51.8	84	28.4
Row 225	Cohort 1	81.7	30.6	49.9	80.7	33.9
Row 226	Cohort 2	81.1	35.7	47.8	86.2	35.8
Row 227	Cohort 3	59	50.6	45.6	56	42.8
Row 228	Cohort 4	71.4	28.7	56.2	79.8	39.5
Row 229	Cohort 1	57.2	34.2	49.3	61.5	49.2
Row 230	Cohort 2	74.7	26.3	50.1	86	32.3
Row 231	Cohort 3	55.3	42.8	47.2	60.8	45.1
Row 232	Cohort 4	62.8	47.3	53.8	66	32.7
Row 233	Cohort 1	69.1	42.4	56.7	58.7	33.1
Row 234	Cohort 2	71.9	38.9	56.2	62.6	49
Row 235	Cohort 3	66.9	45.5	54.7	74.5	41.3
Row 236	Cohort 4	55.2	46.7	55.8	42.6	45.4
Row 237	Cohort 1	67.6	38.2	48.3	79.5	33.3
Row 238	Cohort 2	73.5	34.4	61.2	81.5	40.3
Row 239	Cohort 3	56.3	39.3	47.6	47.1	42.9
Row 240	Cohort 4	65.3	37.9	45.4	57.9	51.7
Row 241	Cohort 1	42.5	45.8	52.2	43.5	48
Row 242	Cohort 2	63.2	53.9	53.3	63.5	46.7
Row 243	Cohort 3	54.3	38.9	48.1	64.3	44.3
Row 244	Cohort 4	56.4	43.2	53.4	56.1	47.3
Row 245	Cohort 1	57.5	44.8	52	57.6	57.3
Row 246	Cohort 2	88.3	26.3	47.4	88.6	37
Row 247	Cohort 3	80.6	26.4	48.9	92.6	24.6
Row 248	Cohort 4	71.2	45.7	52.2	65.4	45.2
Row 249	Cohort 1	81.6	30.4	62.2	69.2	28.2
Row 250	Cohort 2	87.3	21.1	58.4	82.1	31.1

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
Feature 01 vs Feature 02	-0.70	250
Feature 01 vs Feature 03	+0.23	250
Feature 01 vs Feature 04	+0.83	250
Feature 01 vs Feature 05	-0.70	250
Feature 02 vs Feature 03	-0.15	250
Feature 02 vs Feature 04	-0.70	250
Feature 02 vs Feature 05	+0.64	250
Feature 03 vs Feature 04	+0.27	250
Feature 03 vs Feature 05	-0.25	250
Feature 04 vs Feature 05	-0.69	250

upperPanels="mirror" + diagonal="label" · the symmetric SPLOM: the upper triangle redraws the same points transposed, so brushing lights one record up in 12 panels at once. Variables are capped at 4 here too, which makes every panel bigger and easier to read.

Service fleet · five signals crossed

54 services · last 24h · three regions

Strongest linear relationship: p95 latency vs Requests, r = +0.97 on 54 complete pairs. Drawing 4 of 5 variables — 16 panels. Not drawn: Cost per 1k. Point at or tap a mark, or press Tab into the grid and use the arrow keys; Home, End, Page Up and Page Down jump further.

Arrow keys step through records one at a time, Page Up and Page Down move ten, Home and End jump to the first and last. The highlighted record is marked in every panel. All values are in the tables below.

Scatter plot matrix. Service fleet · five signals crossed: 4 variables — p95 latency, Error rate, CPU, Requests — crossed against each other over 54 records, 12 scatter panels with the variable names on the diagonal. Strongest linear relationship: p95 latency and Requests, r = +0.97 over 54 complete pairs. 1 further variable not drawn: Cost per 1k. Every value follows in the data table.
Record	Series	p95 latency (ms)	Error rate (%)	CPU (%)	Requests (/min)
checkout-a	us-east	240 ms	0%	61%	2,811/min
checkout-b	us-east	314 ms	0.05%	66%	3,288/min
checkout-c	us-east	401 ms	0.89%	84%	3,935/min
search-a	us-east	379 ms	0.77%	73%	3,713/min
search-b	us-east	425 ms	1.14%	81%	3,993/min
search-c	us-east	462 ms	1.42%	93%	4,287/min
auth-a	us-east	444 ms	1.29%	83%	4,194/min
auth-b	us-east	60 ms	0.01%	33%	1,413/min
auth-c	us-east	108 ms	0.01%	35%	1,347/min
catalog-a	us-east	87 ms	0.02%	30%	1,285/min
catalog-b	us-east	370 ms	0.74%	73%	4,052/min
catalog-c	us-east	279 ms	0%	64%	3,446/min
cart-a	us-east	107 ms	0%	47%	1,680/min
cart-b	us-east	182 ms	0%	61%	2,812/min
cart-c	us-east	224 ms	0.12%	56%	3,106/min
payments-a	us-east	209 ms	0.06%	56%	2,900/min
payments-b	us-east	294 ms	0.09%	66%	3,260/min
payments-c	us-east	329 ms	0.16%	66%	3,470/min
shipping-a	us-east	98 ms	0.02%	47%	1,788/min
shipping-b	us-east	76 ms	0%	23%	905/min
shipping-c	us-east	276 ms	0.25%	67%	3,462/min
inventory-a	us-east	386 ms	0.82%	85%	3,692/min
inventory-b	eu-west	104 ms	0.03%	46%	1,893/min
inventory-c	eu-west	357 ms	0.47%	72%	3,524/min
recommend-a	eu-west	77 ms	0.07%	28%	1,037/min
recommend-b	eu-west	331 ms	0.48%	81%	3,633/min
recommend-c	eu-west	72 ms	0.03%	35%	1,499/min
reviews-a	eu-west	464 ms	1.27%	91%	4,339/min
reviews-b	eu-west	310 ms	0.3%	75%	3,521/min
reviews-c	eu-west	449 ms	1.4%	88%	4,252/min
media-a	eu-west	197 ms	0.02%	64%	2,782/min
media-b	eu-west	398 ms	1.01%	76%	3,804/min
media-c	eu-west	192 ms	0.07%	56%	2,605/min
notify-a	eu-west	76 ms	0.08%	32%	1,559/min
notify-b	eu-west	449 ms	1.36%	85%	4,123/min
notify-c	eu-west	357 ms	0.69%	82%	3,765/min
billing-a	eu-west	86 ms	0.02%	44%	1,493/min
billing-b	eu-west	496 ms	1.6%	88%	4,351/min
billing-c	eu-west	402 ms	0.85%	76%	3,714/min
identity-a	eu-west	208 ms	0%	61%	2,732/min
identity-b	ap-south	244 ms	0.14%	68%	2,819/min
identity-c	ap-south	263 ms	0.08%	61%	3,015/min
pricing-a	ap-south	34 ms	0.07%	33%	765/min
pricing-b	ap-south	397 ms	0.76%	78%	3,651/min
pricing-c	ap-south	200 ms	0.09%	57%	2,823/min
tracking-a	ap-south	291 ms	0.26%	71%	3,314/min
tracking-b	ap-south	451 ms	1.13%	84%	4,168/min
tracking-c	ap-south	117 ms	0.14%	48%	1,954/min
support-a	ap-south	336 ms	0.37%	66%	3,511/min
support-b	ap-south	344 ms	0.59%	77%	3,872/min
support-c	ap-south	78 ms	0.03%	38%	1,040/min
reports-a	ap-south	247 ms	0.02%	57%	2,616/min
reports-b	ap-south	31 ms	0.05%	33%	782/min
reports-c	ap-south	377 ms	0.9%	80%	3,774/min

Pearson correlation for every variable pair, computed on complete pairs only. It measures straight-line association: a strong curve can score near zero, and one outlier can carry a high score on its own.
Pair	r	Complete pairs
p95 latency vs Error rate	+0.85	54
p95 latency vs CPU	+0.96	54
p95 latency vs Requests	+0.97	54
Error rate vs CPU	+0.80	54
Error rate vs Requests	+0.76	54
CPU vs Requests	+0.97	54