Streaming Data

In this tutorial, you’ll learn how to stream new data in a Vega-Lite visualization. If you are not familiar with Vega-Lite, please read the getting started tutorial first.

We will be using the View API from the Vega, where we update data via the change method. Data in Vega and Vega-Lite is updated either via interactions, or by specifying the data changes directly. In both cases, the data is pushed through the Vega runtime for optimized performance. This tutorial focuses on the latter for streaming data, where no user interaction is involved.

To use view.change, you have to first specify the name of the data source you are updating: view.change('data', ... ). The name needs to be specified in the original spec, as described here. As the second parameter to the view.change method, you should specify what data to add, and what data to remove. To add data, simply supply the array of data entries matching previous schemas, for the example here, it is:

{
  "x": number,
  "y": number,
  "category": number
}

As for removing data points, you could either supply the actual data points to be removed similar to the insert above (which requires your storing the data), or you could filter existing rendered data by some attributes and pass the predicate to the remove() function, such as in this case function (t) { return t.x < minimumX; }. If you are familiar with D3, this insert and remove pattern is similar to the enter exit pattern.

Putting it together, we have:

var changeSet = vega
  .changeset()
  .insert(valueGenerator())
  .remove(function(t) {
    return t.x < minimumX;
  });
view.change('table', changeSet).run();

The view is given to us by the embed method, which helps load Vega/Vega-Lite specs on your webpages. When embed completes successfully, embed returns a fulfilled promise with an object containing the view, which is how we could access the view.change mentioned earlier, and spec, which is the compiled Vega spec from the original Vega-Lite spec given. When embed fails, it returns a rejected promise with an error object.

For simplicity, we have generated some data to simulate real time data updates, but the code could easily be swapped with a server data request. In the example, the “real time” is being simulated by window.setInterval that triggers view.change periodically.

Below is the JavaScript code to run this example. Make sure your html contains a div with id 'chart'.

var vlSpec = {
  $schema: 'https://vega.github.io/schema/vega-lite/v3.json',
  data: {name: 'table'},
  width: 400,
  mark: 'line',
  encoding: {
    x: {field: 'x', type: 'quantitative', scale: {zero: false}},
    y: {field: 'y', type: 'quantitative'},
    color: {field: 'category', type: 'nominal'}
  }
};
vegaEmbed('#chart', vlSpec).then(function(res) {
  /**
   * Generates a new tuple with random walk.
   */
  function newGenerator() {
    var counter = -1;
    var previousY = [5, 5, 5, 5];
    return function() {
      counter++;
      var newVals = previousY.map(function(v, c) {
        return {
          x: counter,
          y: v + Math.round(Math.random() * 10 - c * 3),
          category: c
        };
      });
      previousY = newVals.map(function(v) {
        return v.y;
      });
      return newVals;
    };
  }

  var valueGenerator = newGenerator();
  var minimumX = -100;
  window.setInterval(function() {
    minimumX++;
    var changeSet = vega
      .changeset()
      .insert(valueGenerator())
      .remove(function(t) {
        return t.x < minimumX;
      });
    res.view.change('table', changeSet).run();
  }, 1000);
});

This is the end of this tutorial where you learned how to stream new data into your chart. If you want to use Vega-Lite with websockets, check out the Vega-Lite with websockets demo. You can find more visualizations in the gallery. If you want to further customize your charts, please read the documentation.