Runs a report: either a definition assembled in the constructor, or a saved view by id. The period may be overridden — it belongs to the dashboard, not to the view, so the same report is looked at over a week and over a month. Norms and direction travel with the data so the client colours deviations from the answer instead of hardcoding thresholds, and rows built on fewer than five conversations come back flagged: on such a sample the spread is wider than any deviation read into it.

{
  "columns": [
    {
      "key": "string",
      "title": "string",
      "is_dimension": true,
      "value_type": "Number",
      "aggregation": "Count",
      "norm_min": 0,
      "norm_max": 0,
      "direction": "None",
      "is_date_time": true,
      "dimension_field": "Conversation"
    }
  ],
  "rows": [
    {
      "cells": [
        {
          "column_key": "string",
          "number": 0,
          "text": "string"
        }
      ],
      "conversations": 0,
      "low_sample": true,
      "entity_id": "8161163a-f227-466f-bc01-090a01e80165"
    }
  ],
  "total_conversations": 0,
  "from": "2019-08-24T14:15:22Z",
  "to": "2019-08-24T14:15:22Z",
  "visualization": "Table",
  "warning": "string"
}