Statistical Presets Filter

Statistical presets turn common analytical questions into one-click filters. Use them for metrics where users frequently ask for the top 10%, values above average, the bottom quartile, two-sigma outliers, or negative values.

Source code

grid.plugins = [AdvanceFilterPlugin];
grid.columns = [
  {
    prop: 'margin',
    name: 'Margin',
    filter: ['number', FILTER_STATISTICAL_PRESETS],
  },
];
grid.filter = true;

return [
  { margin: -12 },
  { margin: -3 },
  { margin: 0 },
  { margin: 4 },
  { margin: 7 },
  { margin: 11 },
  { margin: 18 },
  { margin: 25 },
  { margin: 40 },
  { margin: 95 },
];

Quick start

1
import { defineCustomElements } from '@revolist/revogrid/loader';
2
import {
3
  AdvanceFilterPlugin,
4
  FILTER_STATISTICAL_PRESETS,
5
} from '@revolist/revogrid-pro';
6
import '@revolist/revogrid-pro/dist/revogrid-pro.css';
7

8
defineCustomElements();
9

10
const grid = document.createElement('revo-grid');
11
grid.plugins = [AdvanceFilterPlugin];
12
grid.columns = [
13
  { prop: 'team', name: 'Team' },
14
  {
15
    prop: 'margin',
16
    name: 'Margin',
17
    filter: ['number', FILTER_STATISTICAL_PRESETS],
18
  },
19
];
20
grid.filter = true;
21
document.querySelector('#app')!.append(grid);
22
grid.source = [
23
  { team: 'North', margin: 18 },
24
  { team: 'South', margin: -4 },
25
  { team: 'West', margin: 31 },
26
  { team: 'East', margin: 9 },
27
];

Operator and state

Type ID: FILTER_STATISTICAL_PRESETS ('statisticalPresets')
Operator ID: STATISTICAL_PRESET_OPERATOR ('statisticalPreset')
Presets: top10Percent, aboveAverage, bottomQuartile, outliersTwoSigma, and negativeOnly
Value: { preset, lower, upper, lowerInclusive, upperInclusive, outside }

The popup computes statistics from the current finite parsed values. When a preset is selected, the resolved numeric boundaries are serialized into the condition. Those boundaries stay stable when the source later changes; reopen and select the preset again to recompute it.

Restore a preset programmatically

Restore the resolved boundaries, not only the preset name:

1
import {
2
  STATISTICAL_PRESET_OPERATOR,
3
  defineAdvancedFilterConfig,
4
} from '@revolist/revogrid-pro';
5

6
grid.filter = defineAdvancedFilterConfig({
7
  multiFilterItems: {
8
    margin: [
9
      {
10
        id: 1,
11
        type: STATISTICAL_PRESET_OPERATOR,
12
        value: {
13
          preset: 'negativeOnly',
14
          lower: null,
15
          upper: 0,
16
          lowerInclusive: false,
17
          upperInclusive: false,
18
          outside: false,
19
        },
20
        relation: 'and',
21
      },
22
    ],
23
  },
24
});

null represents an unbounded side. outside: true matches values outside the serialized lower and upper boundaries and is used by the two-sigma outlier preset.

Configuration

1
import {
2
  createStatisticalPresetsStructuredFilterType,
3
  defineAdvancedFilterConfig,
4
} from '@revolist/revogrid-pro';
5

6
const presets = createStatisticalPresetsStructuredFilterType({
7
  formatValue: (value) => `${value.toFixed(1)}%`,
8
  formatMatchCount: (count) => `${count.toLocaleString()} matches`,
9
});
10

11
grid.filter = defineAdvancedFilterConfig({
12
  structuredFilterTypes: [presets],
13
});

Both callbacks change presentation only. Use cellParser when the stored value must be normalized before statistics and comparisons are calculated.

Common mistakes

Saving only { preset: 'top10Percent' }; all boundary and inclusion fields are required for a valid restored condition.
Expecting an applied preset to move automatically as rows change; its serialized boundaries are deliberately stable.
Including NaN, Infinity, or nonnumeric strings in expected statistics; only finite parsed values participate.
Treating negativeOnly as “zero or below”; its upper boundary is exclusive, so zero does not match.
Recomputing boundaries independently with different quantile or variance rules when the application can save the condition emitted by the grid.