/documd-visuals

Parallel Coordinates — Multi-Dimensional Risk Screen (ECharts)

Best for: comparing many records across several numeric dimensions when the pattern across dimensions matters more than one ranking Avoid when: the audience is unfamiliar with the dimensions, or only one metric actually drives the decision (use bars/scatter) Answers: which items are extreme, which dimensions trade off against each other, and which profiles cluster together

{
  "color": ["#2b66c4", "#2f9e44", "#f3a33c", "#d1242f", "#7048e8", "#0f9b9b", "#c16f8a", "#7c5a3d"],
  "width": 900,
  "height": 420,
  "title": { "text": "Supplier Risk Screen", "subtext": "Seven suppliers across five screening dimensions", "left": "left" },
  "parallelAxis": [
    { "dim": 0, "name": "Lead time", "min": 0, "max": 40 },
    { "dim": 1, "name": "Defect rate", "min": 0, "max": 10 },
    { "dim": 2, "name": "Single-source", "min": 0, "max": 10 },
    { "dim": 3, "name": "Margin", "min": 0, "max": 40 },
    { "dim": 4, "name": "Recovery days", "min": 0, "max": 30 }
  ],
  "parallel": {
    "left": 56,
    "right": 36,
    "top": 96,
    "bottom": 40,
    "parallelAxisDefault": {
      "nameLocation": "end",
      "nameGap": 18,
      "nameTextStyle": { "fontSize": 11 }
    }
  },
  "series": [
    {
      "type": "parallel",
      "lineStyle": { "width": 2, "opacity": 0.5 },
      "data": [
        { "name": "Supplier A", "value": [18, 2, 4, 24, 9] },
        { "name": "Supplier B", "value": [26, 3, 8, 20, 12] },
        { "name": "Supplier C", "value": [12, 1, 2, 32, 5] },
        { "name": "Supplier D", "value": [34, 6, 9, 16, 18] },
        { "name": "Supplier E", "value": [20, 4, 5, 28, 10] },
        { "name": "Supplier F", "value": [28, 7, 7, 18, 21] },
        { "name": "Supplier G", "value": [10, 2, 3, 35, 6] }
      ]
    }
  ]
}

Key Options

Data Shape

Each record is one array of values in the exact order of the parallelAxis dimensions. This chart assumes all dimensions are numeric and comparable in the sense of “more/less along an axis”.

Pitfalls

• ❌ Too many records at once → ✅ parallel coordinates saturate quickly; keep it to a small screened set in report exports
• ❌ Unclear axis ranges → ✅ always set min and max, or the chart becomes hard to compare across exports
• ❌ Using it for categorical dimensions only → ✅ this view is strongest when several numeric measures compete
• ❌ Expecting readers to infer labels on hover → ✅ static exports need a small enough record set that the overall shapes remain readable without interaction

Alternatives