Search code examples
powerbipowerbi-desktopvega-litedeneb

Vega-Lite (Deneb): Donut Chart - Unable to aggregate text marks against aggregated categories

Any assitance warmly welcomed. I imagine I'm missing something obvious, but very stuck with this one.

Please see images, and links to Vega-Lite editors below:

I have two fields "Medium" (aka Subcategory) and "Medium-Type" (aka Category). I'm able to produce a donut by subcategory, with corresponding data labels. However, when aggregating the donut by Category, the donut segments behave correctly, but I'm unable to aggregate the text marks in the same way.

Dataset:

Dataset

EG1: Donut - SubCategory (VL Editor) enter image description here

EG2: Donut - Category (VL Editor) enter image description here

Thanks community!

Solution

  • Try and update your transform like:

    {
"joinaggregate": [{"op": "sum", "field": "Advertising Share", "as": "calc"}],
"groupby": ["Medium-Type"]
}

    Then replace all "Advertising Share" with calc. Also remove "aggregate": "sum" from theta since it is already aggregated.

    {
  "data": {
    "values": [
      {
        "Medium": "Direct Mail",
        "Advertising Share": 3.2,
        "Medium-Type": "Hard Copy"
      },
      {
        "Medium": "Posters",
        "Advertising Share": 3.4,
        "Medium-Type": "Hard Copy"
      },
      {"Medium": "Press", "Advertising Share": 5.4, "Medium-Type": "Hard Copy"},
      {
        "Medium": "Television",
        "Advertising Share": 69.8,
        "Medium-Type": "Digital"
      },
      {"Medium": "Cinema", "Advertising Share": 0.7, "Medium-Type": "Digital"},
      {"Medium": "Radio", "Advertising Share": 2.1, "Medium-Type": "Radio"},
      {
        "Medium": "Internet",
        "Advertising Share": 15.4,
        "Medium-Type": "Digital"
      }
    ]
  },
  "transform": [
    {
      "joinaggregate": [
        {"op": "sum", "field": "Advertising Share", "as": "calc"}
      ],
      "groupby": ["Medium-Type"]
    }
  ],
  "width": "500",
  "height": "500",
  "padding": {"top": 5, "right": 0, "bottom": 5, "left": 0},
  "layer": [
    {
      "mark": {
        "type": "arc",
        "outerRadius": {"expr": "(height * 0.5 ) "},
        "innerRadius": {"expr": "(height * 0.3 ) "},
        "stroke": "black",
        "cornerRadius": 5,
        "padAngle": 0.04
      }
    },
    {
      "mark": {
        "type": "text",
        "fontSize": 20,
        "fontWeight": "bold",
        "radius": {"expr": "(height * 0.55)"}
      },
      "encoding": {
        "text": {"field": "calc", "type": "quantitative"},
        "theta": {"field": "calc", "type": "quantitative"},
        "color": {"value": "black"}
      }
    }
  ],
  "encoding": {
    "theta": {
      "field": "calc",
      "type": "quantitative",
      "stack": "normalize"
    },
    "order": {
      "field": "calc",
      "type": "quantitative",
      "aggregate": "sum",
      "sort": "descending"
    },
    "color": {
      "legend": {
        "title": "Medium (%)",
        "legendY": -10,
        "orient": "right",
        "symbolSize": 450,
        "labelFontSize": 25,
        "titleFontSize": 30
      },
      "field": "Medium-Type",
      "type": "nominal",
      "sort": {"field": "calc", "order": "descending"}
    },
    "opacity": {"value": 0.9}
  },
  "view": {"stroke": null}
}

    In other words you were very, very close.

    Adam