Stacked bar chart is created even though aggregation function is non-summative

Question

I have a concatenated graph which features a main line graph which has a brush selection tool allowing the user to pan across the lines and points and change the data on 4 other graphs. For one of the other graphs, I have attempted to take the average of line graph data but it doesn't work. Instead of giving me a singular bar, I get stacked bars and the error: "Stacking is applied even though the aggregate function is non-summative ("mean")".

Here is my code:

{
  "$schema": "https://vega.github.io/schema/vega-lite/v4.json",
  "title": "This is kinda sick yo",
  "data": {
    "url": "data/test3.csv"
  },
  "hconcat": [
    {
      "encoding": {
        "color": {
          "condition": {
            "selection": "brush",
            "title": "Species",
            "field": "Species",
            "type": "nominal",
            "scale": {"range": ["green", "#FFFF00", "red"]}
          },
          "value": "lightgray"
        },
        "x": {
          "field": "Variable",
          "type": "nominal",
          "axis": {"labelAngle": -45, "title": "Element",
        "grid": false}
        },
        "y": {
          "title": "Total",
          "field": "Total",
          "type": "quantitative"
        },
          "tooltip": [
      {"field": "Variable", "type": "nominal"},
      {"field": "Total", "type": "quantitative"},
    ]
      },
      "width": 550,
      "height": 300,
      "mark": {"type": "line", "point": "true"},
      "selection": {"brush": {"encodings": ["x"], "type": "interval"}},
      "transform": [{"filter": {"selection": "click"}}]
    },
      {
      "encoding": {
        "color": {
          "condition": {
            "selection": "click",
            "field": "Total",
            "type": "quantitative",
            "scale": {"range": ["green", "#FFFF00", "red"]}
          },
          "value": "lightgray"
        },
        "y": {"field": "Total", "aggregate": "average"},
        "x": {"title": "Species", "field": "Species", "type": "nominal"},
          "tooltip": [
      {"field": "Species", "type": "nominal"},
      {"field": "Total", "type": "quantitative", "aggregate": "average"},
      {"field": "Variable", "type": "nominal"}
    ]
      },
      "height": 300,
      "width": 80,
      "mark": "bar",
      "selection": {"click": {"encodings": ["color"], "type": "multi"}},
      "transform": [{"filter": {"selection": "brush"}}, ]
    },
    {
      "encoding": {
        "color": {
          "condition": {
            "selection": "click",
            "field": "Sex",
            "type": "nominal",
            "scale": {"range": ["#993162", "#75b0a2", "grey"]},
            "legend": null
          },
          "value": "lightgray"
        },
        "y": {"field": "Fisher Sex Value", "type": "quantitative", "aggregate": "mean"},
        "x": {"title": "Sex", "field": "Sex", "type": "nominal"},
          "tooltip": [
      {"field": "Sex", "type": "nominal"},
      {"field": "Fisher Sex Value", "type": "quantitative", "aggregate": "mean"},
    ]
      },
      "height": 300,
      "width": 75,
      "mark": "bar",
      "selection": {"click": {"encodings": ["color"], "type": "multi"}},
      "transform": [{"filter": {"selection": "brush"}}]
    },
      {
      "encoding": {
        "color": {
          "condition": {
            "selection": "click",
            "field": "Sex",
            "type": "nominal",
            "scale": {"range": ["#993162", "#75b0a2", "grey"]},
            "legend": null
          },
          "value": "lightgray"
        },
        "y": {"field": "Mink Sex Value", "type": "quantitative", "aggregate": "mean"},
        "x": {"title": "Sex", "field": "Sex", "type": "nominal"},
          "tooltip": [
      {"field": "Sex", "type": "nominal"},
      {"field": "Mink Sex Value", "type": "quantitative", "aggregate": "mean"},
    ]
      },
      "height": 300,
      "width": 75,
      "mark": "bar",
      "selection": {"click": {"encodings": ["color"], "type": "multi"}},
      "transform": [{"filter": {"selection": "brush"}}]
    },
      {
      "encoding": {
        "color": {
          "condition": {
            "selection": "click",
            "field": "Sex",
            "type": "nominal",
            "scale": {"range": ["#993162", "#75b0a2", "grey"]}
          },
          "value": "lightgray"
        },
        "y": {"field": "Otter Sex Value", "type": "quantitative", "aggregate": "mean"},
        "x": {"title": "Sex", "field": "Sex", "type": "nominal"},
          "tooltip": [
      {"field": "Sex", "type": "nominal"},
      {"field": "Otter Sex Value", "type": "quantitative", "aggregate": "mean"},
    ]
      },
      "height": 300,
      "width": 75,
      "mark": "bar",
      "selection": {"click": {"encodings": ["color"], "type": "multi"}},
      "transform": [{"filter": {"selection": "brush"}}]
    }
  ]
}

The first graph is the line graph and the second graph is the one where aggregation fails and I get stacks.Here is an image of what the graph looks like currently. Any help would be much appreciated.

Answer 1

Vega-Lite's encoding aggregations will implicitly group by unaggregated fields you specify in a set of encodings. A simplified version of the second chart's encoding looks like this:

{
  "encoding": {
  "color": {"field": "Total"},
  "y": {"field": "Total", "aggregate": "average"},
  "x": {"field": "Species"},
  "tooltip": [
    {"field": "Species"},
    {"field": "Total", "aggregate": "average"},
    {"field": "Variable"}
  ]

The unaggregated encodings are ["Total", "Species", "Variable"], so the operation will group-by these before computing the average of Total within each group. Grouping by unique values of Total before taking the mean of Total in each group is probably not what you were hoping for.

Perhaps removing the color encoding from that chart will give you more meaningful results.