CH-DST-01 . family DST
HISTOGRAM
contiguous columns counting values per equal bin. Needs:
one numeric measure binned into equal intervals with counts.
distribution
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
- showing the shape of one distribution: skew, mode, outliers
- first look at any new numeric measure
Do not use when
- comparing several distributions (CH-DST-04 or CH-DST-09)
- categorical counts (CH-RNK-01; bins must be numeric intervals)
See instead
- the question is what share clears a threshold: CH-DST-06
- comparing shapes across groups: CH-DST-09
Alternatives
- CH-DST-06 (service-level reading)
- CH-DST-09 (smooth shape)
Rules
- bars touch: 1px seam only, the variable is continuous
- equal bin widths, edges labeled at the boundaries, not centers
- counts on the y axis start at zero
Never do this
- gaps between bars
- unequal bins without width-proportional areas
- labeling bin centers
- more than 20 bins
QA before delivering
- bars contiguous
- edges ascending and labeled
- y starts at zero
- unit once
- source present
Data shape
| title | action title with a number |
| source | bottom-left always |
| unit | on the top count tick |
| edges | n+1 ascending bin edges, labeled under the boundaries |
| counts | n non-negative counts |
Sample data (golden)
{
"title": "Two thirds of lots clear customs within 15 days",
"eyebrow": "CLEARANCE TIME",
"source": "Source: customs log, H1",
"unit": "lots",
"edges": [
0,
5,
10,
15,
20,
25,
30,
35,
40
],
"counts": [
18,
42,
31,
17,
9,
5,
3,
1
]
}