PROVEGA: Progressive Vega-Lite Extension

PROVEGA is an extension of Vega-Lite aimed at supporting progressive visualization. It introduces new properties under the provega namespace.

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "url": "data/large_dataset.csv"
  },
  "mark": "line",
  "encoding": {
    "x": { "field": "timestamp", "type": "temporal" },
    "y": { "field": "value", "type": "quantitative" }
  },
  "provega": {
    "progression": {
      "chunking": {
      },
      "control": {
      },
      "monitoring": {
      }
    }
  }
}

🧭 Table of Contents

• 1. Progression
  • 1.1 Chunking
  • 1.2 Control
  • 1.3 Monitoring
  • 1.4 Quality
• 2. Contribute

1. Progression

1.1 chunking

Defines how the data is chunked and read progressively. The type field is mandatory.

The reading field is used when the dataset is passed as a single CSV file (or JSON array) embedded in the spec, or as URL endpoint, and must be sampled and chunked progressively during rendering.

Alternatively, if the Vega-Lite data.values field already contains an array of arrays (each sub-array representing a pre-defined chunk), then only the reading.frequency value will be considered to control playback speed across these chunks.

📄 Simple Example Making Data Chunking from a Dataset

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values": [
      { "x": 1, "y": 5 }, 
      { "x": 2, "y": 10 },
      { "x": 3, "y": 15 }, 
      { "x": 4, "y": 20 },
      { "x": 5, "y": 25 },
      { "x": 6, "y": 30 },
      { "x": 7, "y": 35 },
      { "x": 8, "y": 40 },
      { "x": 9, "y": 45 },
      { "x": 10, "y": 50 }
    ]
  },
  "mark": "line",
  "encoding": {
    "x": { "field": "x", "type": "quantitative" },
    "y": { "field": "y", "type": "quantitative" }
  },
  "provega": {
    "progression": {
      "chunking": {
        "type": "data",
        "reading": {
          "frequency": 1000,
          "chunk_size": 2,
          "method": "sequential",
          "ascending": true
        }
      }
    }
  }
}

📄 Simple Example Using Pre-Chunks

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values": [
      [ { "x": 1, "y": 5 }, { "x": 2, "y": 10 } ],
      [ { "x": 3, "y": 15 }, { "x": 4, "y": 20 } ],
      [ { "x": 5, "y": 25 } ]
    ]
  },
  "mark": "line",
  "encoding": {
    "x": { "field": "x", "type": "quantitative" },
    "y": { "field": "y", "type": "quantitative" }
  },
  "provega": {
    "progression": {
      "chunking": {
        "type": "data",
        "reading": {
          "frequency": 1000
        }
      }
    }
  }
}

1.2 control

Defines the control modes and interactions with the progressive process.

1.3 monitoring

Adds monitoring tools to track progress and quality of progressive visualization.

1.3.1 change.area

It is possible to customize the appeareance of the change area. Instead of passing a sigle bool value to enable the change area, pass an object.

1.3.2 change.mark

It is possible to customize the appeareance of the changed marks. Instead of passing a sigle bool value to enable the highlight, pass an object.

1.4 quality

Tracks various indicators of progressive quality. Each field is optional and accepts a variable name (string) or true for defaults.

  "quality": {
          "absolute_progress": {
            "on_data_input": true,
            "on_result_output": true,
            "on_visual_output": true
          },
          "relative_progress": {
            "on_data_input": true,
            "on_result_output": true,
            "on_visual_output": true
          },
          "relative_stability": {
            "on_data_input": true,
            "on_result_output": true,
            "on_visual_output": true
          },
          "absolute_certainty": {
            "on_result_output": true
          }
    }
 

📄 Example: Vega-Lite Spec with PROVEGA

This example demonstrates how to use provega properties in a Vega-Lite specification to enable progressive chunking and monitoring:

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "url": "data/large_dataset.csv"
  },
  "mark": "line",
  "encoding": {
    "x": { "field": "timestamp", "type": "temporal" },
    "y": { "field": "value", "type": "quantitative" }
  },
  "provega": {
    "progression": {
      "chunking": {
        "type": "data",
        "reading": {
          "method": "sequential",
          "chunk_size": 20,
          "frequency": 500
        }
      },
      "control": {
        "mode": "exploration",
        "pause": true,
        "stop": true,
        "backward": true,
        "min_frequency": 300
      },
      "monitoring": {
        "aliveness": true,
        "progressbar": true,
        "change": {
          "area": true,
          "noise_reduction": true
        },
        "quality": {
          "absolute_progress": {
            "on_data_input": true
          }
        }
      }
    }
  }
}

🤝 Contribute: help us extend ProVega

The grammar is open, and so is everything around it. A chunking strategy, a monitoring indicator, a control the literature uses and provega does not express yet: fork the repository, add it, and send a pull request. Bug reports and examples that break the grammar are just as welcome.

github.com/XAIber-lab/provega