Visualize.Contour.Density (Visualize v0.2.35)

Copy Markdown View Source

Kernel density estimation for contour generation from point data.

Estimates a continuous density surface from scattered points using Gaussian kernel density estimation, then generates contours from the resulting grid.

Examples

# Generate density contours from points
points = [
  %{x: 10, y: 20},
  %{x: 15, y: 25},
  %{x: 12, y: 22},
  # ... more points
]

density = Visualize.Contour.Density.new()
  |> Visualize.Contour.Density.x(fn d -> d.x end)
  |> Visualize.Contour.Density.y(fn d -> d.y end)
  |> Visualize.Contour.Density.size(100, 100)
  |> Visualize.Contour.Density.bandwidth(10)
  |> Visualize.Contour.Density.thresholds(10)

contours = Visualize.Contour.Density.compute(density, points)

Algorithm

Uses Gaussian kernel density estimation:

  • For each grid cell, sum the weighted contributions from all points
  • Weight is based on distance using Gaussian kernel: K(u) = exp(-u²/2)
  • Bandwidth controls the smoothness (larger = smoother)

Summary

Types

t()

Functions

bandwidth(density, bw)

Sets the bandwidth (kernel size) for density estimation.

cell_size(density, size)

Sets the cell size for the density grid

compute(density, points)

Computes density contours from point data.

new()

Creates a new density estimator

render(density, points)

Renders density contours as SVG paths.

size(density, width, height)

Sets the output dimensions

thresholds(density, values)

Sets the threshold values for contour generation.

weight(density, accessor)

Sets the weight accessor function

x(density, accessor)

Sets the x accessor function

y(density, accessor)

Sets the y accessor function

Types

t()

@type t() :: %Visualize.Contour.Density{
  bandwidth: number(),
  cell_size: pos_integer(),
  height: pos_integer(),
  thresholds: [number()] | pos_integer(),
  weight: (any() -> number()) | nil,
  width: pos_integer(),
  x: (any() -> number()) | nil,
  y: (any() -> number()) | nil
}

Functions

bandwidth(density, bw)

@spec bandwidth(t(), number()) :: t()

Sets the bandwidth (kernel size) for density estimation.

Larger values produce smoother results but may obscure detail.

cell_size(density, size)

@spec cell_size(t(), pos_integer()) :: t()

Sets the cell size for the density grid

compute(density, points)

@spec compute(t(), Visualize.Data.Table.source()) :: [
  Visualize.Contour.contour_result()
]

Computes density contours from point data.

Returns a list of contour objects compatible with Visualize.Contour.

new()

@spec new() :: t()

Creates a new density estimator

render(density, points)

@spec render(t(), [any()]) :: [%{value: number(), path: String.t()}]

Renders density contours as SVG paths.

size(density, width, height)

@spec size(t(), pos_integer(), pos_integer()) :: t()

Sets the output dimensions

thresholds(density, values)

@spec thresholds(t(), [number()] | pos_integer()) :: t()

Sets the threshold values for contour generation.

Can be a list of specific values or a count (generates that many levels).

weight(density, accessor)

@spec weight(t(), (any() -> number())) :: t()

Sets the weight accessor function

x(density, accessor)

@spec x(t(), (any() -> number())) :: t()

Sets the x accessor function

y(density, accessor)

@spec y(t(), (any() -> number())) :: t()

Sets the y accessor function