Specs

concat(t(), [t()], :wrappable | :horizontal | :vertical) :: t()

Builds a concatenated multi-view specification from the given list of single view specifications.

The concat type must be either :wrappable (default), :horizontal or :vertical.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.concat([
  Vl.new()
  |> ...,
  Vl.new()
  |> ...,
  Vl.new()
  |> ...
])

Vl.new()
|> Vl.data_from_values(...)
|> Vl.concat(
  [
    Vl.new()
    |> ...,
    Vl.new()
    |> ...
  ],
  :horizontal
)

See the docs for more details.

Specs

config(t(), keyword()) :: t()

Adds view configuration to the specification.

Configuration allows for setting general properties of the visualization.

All provided options are converted to configuration properties and merged with the existing configuration in a shallow manner.

Examples

Vl.new()
|> ...
|> Vl.config(
  view: [stroke: :transparent],
  padding: 100,
  background: "#333333"
)

See the docs for more details.

Specs

data(t(), keyword()) :: t()

Sets data properties in the specification.

Defining the data source is usually the first step when building a graphic. For most use cases it's preferable to use more specific functions like data_from_url/3, data_from_values/3, or data_from_series/3.

All provided options are converted to data properties.

Examples

Vl.new()
|> Vl.data(sequence: [start: 0, stop: 12.7, step: 0.1, as: "x"])
|> ...

See the docs for more details.

Specs

data_from_series(t(), Enumerable.t(), keyword()) :: t()

Sets inline data in the specification.

This is an alternative to data_from_values/3, useful when you have a separate list of values for each data column.

Examples

xs = 1..100
ys = 1..100

Vl.new()
|> Vl.data_from_series(x: xs, y: ys)
|> ...

Specs

data_from_url(t(), String.t(), keyword()) :: t()

Sets data URL in the specification.

The URL should be accessible by whichever client renders the specification, so preferably an absolute one.

All provided options are converted to data properties.

Examples

Vl.new()
|> Vl.data_from_url("https://vega.github.io/editor/data/penguins.json")
|> ...

Vl.new()
|> Vl.data_from_url("https://vega.github.io/editor/data/stocks.csv", format: :csv)
|> ...

See the docs for more details.

Specs

data_from_values(t(), Enumerable.t(), keyword()) :: t()

Sets inline data in the specification.

values should be an enumerable of data records, where each record is a key-value structure.

All provided options are converted to data properties.

Examples

data = [
  %{"category" => "A", "score" => 28},
  %{"category" => "B", "score" => 55}
]

Vl.new()
|> Vl.data_from_values(data)
|> ...

See the docs for more details.

Specs

encode(t(), atom(), keyword()) :: t()

Adds an encoding entry to the specification.

Visual channel represents a property of a visual mark, for instance the :x and :y channels specify where a point should be placed. Encoding defines the source of values for those channels.

In most cases you want to map specific data field to visual channels, prefer the encode_field/4 function for that.

All provided options are converted to channel properties.

Examples

Vl.new()
|> Vl.encode(:x, value: 2)
|> ...

Vl.new()
|> Vl.encode(:y, aggregate: :count, type: :quantitative)
|> ...

Vl.new()
|> Vl.encode(:y, field: "price")
|> ...

See the docs for more details.

Specs

encode_field(t(), atom(), String.t(), keyword()) :: t()

Adds field encoding entry to the specification.

A shorthand for encode/3, mapping a data field to a visual channel.

For example, if the data has "price" and "time" fields, you could map "time" to the :x channel and "price" to the :y channel. This, combined with a line mark, would then result in price-over-time plot.

All provided options are converted to channel properties.

Types

Field data type is automatically inferred, but oftentimes needs to be specified explicitly to get the desired result. The :type option can be either of:

• :quantitative - when the field expresses some kind of quantity, typically numerical

• :temporal - when the field represents a point in time

• :nominal - when the field represents a category

• :ordinal - when the field represents a ranked order. It is similar to :nominal, but there is a clear order of values

• :geojson - when the field represents a geographic shape adhering to the GeoJSON specification

See the docs for more details on types.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.mark(:point)
|> Vl.encode_field(:x, "time", type: :temporal)
|> Vl.encode_field(:y, "price", type: :quantitative)
|> Vl.encode_field(:color, "country", type: :nominal)
|> Vl.encode_field(:size, "count", type: :quantitative)
|> ...

Vl.new()
|> Vl.encode_field(:x, "date", time_unit: :month, title: "Month")
|> Vl.encode_field(:y, "price", type: :quantitative, aggregate: :mean, title: "Mean product price")
|> ...

See the docs for more details.

Specs

encode_repeat(t(), atom(), :repeat | :row | :column | :layer, keyword()) :: t()

Adds repeated field encoding entry to the specification.

A shorthand for encode/3, mapping a field to a visual channel, as given by the repeat operator.

Repeat type must be either :repeat, :row, :column or :layer and correspond to the repeat definition.

All provided options are converted to channel properties.

Examples

See repeat/3 to see the full picture.

See the docs for more details.

Specs

facet(t(), keyword(), t()) :: t()

Builds a facet multi-view specification from the given single-view template.

Facet definition must be either a field definition or a row/column mapping.

Note that you can also create facet graphics by using the :facet, :column and :row encoding channels.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.facet(
  [field: "country"],
  Vl.new()
  |> Vl.mark(:bar)
  |> Vl.encode_field(:x, ...)
  |> Vl.encode_field(:y, ...)
)

Vl.new()
|> Vl.data_from_values(...)
|> Vl.facet(
  [
    row: [field: "country", title: "Country"],
    column: [field: "year", title: "Year"]
  ]
  Vl.new()
  |> Vl.mark(:bar)
  |> Vl.encode_field(:x, ...)
  |> Vl.encode_field(:y, ...)
)

See the docs for more details.

Specs

from_json(String.t()) :: t()

Parses the given Vega-Lite JSON specification and wraps in the VegaLite struct for further processing.

Examples

Vl.from_json("""
{
  "data": { "url": "https://vega.github.io/editor/data/cars.json" },
  "mark": "point",
  "encoding": {
    "x": { "field": "Horsepower", "type": "quantitative" },
    "y": { "field": "Miles_per_Gallon", "type": "quantitative" }
  }
}
""")

See the docs for more details.

Specs

from_spec(spec()) :: t()

Wraps the given Vega-Lite specification in the VegaLite struct for further processing.

There is also from_json/1 that handles JSON parsing for you.

See the docs for more details.

Specs

layers(t(), [t()]) :: t()

Builds a layered multi-view specification from the given list of single view specifications.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.layers([
  Vl.new()
  |> Vl.mark(:line)
  |> Vl.encode_field(:x, ...)
  |> Vl.encode_field(:y, ...),
  Vl.new()
  |> Vl.mark(:rule)
  |> Vl.encode_field(:y, ...)
  |> Vl.encode(:size, value: 2)
])

Vl.new()
|> Vl.data_from_values(...)
# Note: top-level data, encoding, transforms are inherited
# by the child views unless overriden
|> Vl.encode_field(:x, ...)
|> Vl.layers([
  ...
])

See the docs for more details.

Specs

mark(t(), atom(), keyword()) :: t()

Sets mark type in the specification.

Mark is a predefined visual object like a point or a line. Visual properties of the mark are defined by encoding.

All provided options are converted to mark properties.

Examples

Vl.new()
|> Vl.mark(:point)
|> ...

Vl.new()
|> Vl.mark(:point, tooltip: true)
|> ...

See the docs for more details.

Specs

new(keyword()) :: t()

Returns a new specification wrapped in the VegaLite struct.

All provided options are converted to top-level properties of the specification.

Examples

Vl.new(
  title: "My graph",
  width: 200,
  height: 200
)
|> ...

See the docs for more details.

Specs

param(t(), String.t(), keyword()) :: t()

Adds a parameter to the specification.

Parameters are the basic building blocks for introducing interactions to graphics.

All provided options are converted to parameter properties.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.concat([
  Vl.new()
  # Define a parameter named "brush", whose value is a user-selected interval on the x axis
  |> Vl.param("brush", select: [type: :interval, encodings: [:x]])
  |> Vl.mark(:area)
  |> Vl.encode_field(:x, "date", type: :temporal)
  |> ...,
  Vl.new()
  |> Vl.mark(:area)
  # Use the "brush" parameter value to limit the domain of this view
  |> Vl.encode_field(:x, "date", type: :temporal, scale: [domain: [param: "brush"]])
  |> ...
])

See the docs for more details.

Specs

projection(t(), keyword()) :: t()

Adds a projection spec to the specification.

Projection maps longitude and latitude pairs to x, y coordinates.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.projection(type: :albers_usa)
|> Vl.mark(:circle)
|> Vl.encode_field(:longitude, "longitude", type: :quantitative)
|> Vl.encode_field(:latitude, "latitude", type: :quantitative)

See the docs for more details.

Specs

repeat(t(), keyword(), t()) :: t()

Builds a repeated multi-view specification from the given single-view template.

Repeat definition must be either a list of fields or a row/column/layer mapping. Then some channels can be bound to a repeated field using encode_repeat/4.

Examples

# Simple repeat
Vl.new()
|> Vl.data_from_values(...)
|> Vl.repeat(
  ["temp_max", "precipitation", "wind"],
  Vl.new()
  |> Vl.mark(:line)
  |> Vl.encode_field(:x, "date", time_unit: :month)
  # The graphic will be reapeated with :y mapped to "temp_max",
  # "precipitation" and "wind" respectively
  |> Vl.encode_repeat(:y, :repeat, aggregate: :mean)
)

# Grid repeat
Vl.new()
|> Vl.data_from_values(...)
|> Vl.repeat(
  [
    row: [
      "beak_length",
      "beak_depth",
      "flipper_length",
      "body_mass"
    ],
    column: [
      "body_mass",
      "flipper_length",
      "beak_depth",
      "beak_length"
    ]
  ],
  Vl.new()
  |> Vl.mark(:point)
  # The graphic will be repeated for every combination of :x and :y
  # taken from the :row and :column lists above
  |> Vl.encode_repeat(:x, :column, type: :quantitative)
  |> Vl.encode_repeat(:y, :row, type: :quantitative)
)

See the docs for more details.

Specs

resolve(t(), atom(), keyword()) :: t()

Adds a resolve entry to the specification.

Resolution defines how multi-view graphics are combined with regard to scales, axis and legend.

Example

Vl.new()
|> Vl.data_from_values(...)
|> Vl.layers([
  Vl.new()
  |> ...,
  Vl.new()
  |> ...
])
|> Vl.resolve(:scale, y: :independent)

See the docs for more details.

Specs

to_spec(t()) :: spec()

Returns the underlying Vega-Lite specification.

The result is a nested Elixir datastructure that serializes to Vega-Lite JSON specification.

See the docs for more details.

Specs

transform(t(), keyword()) :: t()

Adds a transformation to the specification.

Transformation describes an operation on data, like calculating new fields, aggregating or filtering.

All provided options are converted to transform properties.

Examples

Vl.new()
|> Vl.data_from_values(...)
|> Vl.transform(calculate: "sin(datum.x)", as: "sin_x")
|> ...

Vl.new()
|> Vl.data_from_values(...)
|> Vl.transform(filter: "datum.height > 150")
|> ...

Vl.new()
|> Vl.data_from_values(...)
|> Vl.transform(regression: "price", on: "date")
|> ...

See the docs for more details.