CH-PTW-07 . family PTW
MARIMEKKO
100% stacked columns whose widths carry a second total. Needs:
2-6 categories with a total each, split into 2-4 shared segments.
part-to-wholecompositionmagnitude
DL-01 AGRAW
DL-02 KATA
DL-03 LOGOS
DL-04 SHILPA
DL-05 OBEYA
DL-06 LEX
DL-07 SAGA
DL-08 SENTINEL
DL-09 ATLAS
DL-10 BASIRA
DL-11 SUTRA
DL-12 NOROSHI
DL-13 MIZAN
DL-14 EVIDENTIA
DL-15 ABRID
DL-16 TELOS
Use when
- two composition questions at once: how big is each category, and how does each split
- market or spend maps where width = size and height = share
Do not use when
- audiences reading it for the first time without a walkthrough
- more than 6 columns or 4 segments (it becomes a mosaic of noise)
See instead
Alternatives
- CH-PTW-02 (equal widths)
- CH-PTW-05 (one level, area only)
Rules
- column width proportional to its total; heights are per-column shares
- segment names labeled on the first column, never a legend
- share printed in cells that fit; small cells stay silent
- column totals printed under the labels
Never do this
- more than 6x4 cells
- legends
- 3D mosaic effects
- widths not proportional to totals
QA before delivering
- width = total, height = share
- segments named once
- totals printed
- unit once
- source present
Data shape
| title | action title with a number |
| source | bottom-left always |
| unit | on first column total |
| segment_names | 2-4 names, shared by every column |
| columns | [{label, segments: [values]}], width = column total |
Sample data (golden)
{
"title": "North holds a third of spend and commits the deepest share",
"eyebrow": "SPEND MAP . 2026",
"source": "Source: sample dataset v1, 2026-07-12",
"unit": "k$",
"segment_names": [
"Committed",
"Planned",
"Open"
],
"columns": [
{
"label": "North",
"segments": [
70,
30,
20
]
},
{
"label": "Central",
"segments": [
40,
30,
20
]
},
{
"label": "South",
"segments": [
20,
25,
15
]
},
{
"label": "East",
"segments": [
15,
15,
15
]
},
{
"label": "West",
"segments": [
5,
10,
15
]
}
]
}