Pivot table widgets in Decision Optimization experiments
You can add pivot tables in the Visualization view to display data from an existing (input or output) table aggregated in different ways.
To show data aggregated, select the table columns and rows that you want to display and select the appropriate aggregator to compute the data for each cell.
For example, the JSON code that is generated for a pivot table is as follows:
{
"name": "",
"type": "Pivot",
"props": {
"container": "Pivot2",
"data": "party",
"spec": {
"config": {},
"columns": [
{
"property": "Party Size",
"label": "Party Size",
"type": "Number",
"visible": true,
"width": 0,
"style": {}
}
],
"rows": [
{
"property": "Payer Gender",
"label": "Payer Gender",
"type": "String",
"visible": true,
"width": 0,
"style": {}
},
{
"property": "Day of Week",
"label": "Day of Week",
"type": "String",
"visible": true,
"width": 0,
"style": {}
}
],
"tableRenderer": "Table Row Heatmap",
"aggregator": "Count Unique Values",
"aggregatedCol1": "Tip"
},
"search": ""
}
}
Field | Description |
---|---|
name |
The name of the widget |
type |
Pivot |
container |
The selected scenario |
data |
The name of the existing table to pivot |
columns |
An array of columns of the existing table that are used as pivot column headers |
rows |
An array of rows of the existing table that are used as pivot rows headers |
tableRenderer |
The type of table rendered. Choose from these options Table , Table
Heatmap , Table Col Heatmap , and Table Row Heatmap |
aggregator |
The function that computes the cells from the chosen columns and rows. Several choices are
available:Count , Count Unique Values , List Unique
Value , Sum , Integer Sum , Average ,
Median , Sample variance , Sample Standard
Deviation , Minimum , Maximum , First ,
Last , Sum over Sum , Sum as Fraction of Total ,
Sum as Fraction of Rows , Sum as Fraction of Columns , Count
as Fraction of Total , Count as Fraction of Rows , Count as Fraction
of Columns |
|
The name of the columns needed for those aggregators that require table columns to aggregate
the data.Count aggregators require no such column name. The Sum over
Sum aggregator requires 2 columns, and all other aggregators require one column. |