Plots.jl

Plots.jl and Makie.jl functions mostly work out of the box on AbstractDimArray, although not with the same results - they choose to follow each packages default behaviour as much as possible.

This will plot a line plot with 'a', 'b' and 'c' in the legend, and values 1-10 on the labelled X axis:

Plots.jl support is deprecated, as development is moving to Makie.jl

Makie.jl

Makie.jl functions also mostly work with AbstractDimArray and will permute and reorder axes into the right places, especially if X/Y/Z/Ti dimensions are used.

In Makie a DimMatrix will plot as a heatmap by default, but it will have labels and axes in the right places:

using DimensionalData, CairoMakie

A = rand(X(10:10:100), Y([:a, :b, :c]))
Makie.plot(A; colormap=:inferno)

Other plots also work, here DD ignores the axis order and instead favours the categorical variable for the X axis:

Makie.rainclouds(A)

AlgebraOfGraphics.jl

AlgebraOfGraphics.jl is a high-level plotting library built on top of Makie.jl that provides a declarative algebra for creating complex visualizations, similar to ggplot2's "grammar of graphics" in R. It allows you to construct plots using algebraic operations like * and +, making it easy to create sophisticated graphics with minimal code.

Any DimensionalArray is also a Tables.jl table, so it can be used with AlgebraOfGraphics.jl directly. You can indicate columns in mapping with Symbols directly (like :X or :Y), or you can use the Dim type directly (like X or Y)!

Note

If your dimensional array is not named, then you can access the data as the :unnamed column. Otherwise, the data is accessible by its name.

Let's start with a simple example, and plot a 2-D dimarray as a scatter plot, colored by its value.

using DimensionalData, AlgebraOfGraphics, CairoMakie

A = DimArray(rand(10, 10), (X(1:10), Y(1:10)), name = :data)

data(A) * mapping(X, Y; color = :data) * visual(Scatter) |> draw

Don't restrict yourself to standard visualizations! You can use all of AlgebraOfGraphics' features.

Let's plot each X-slice, faceted in Y:

data(A) * mapping(X, :data; layout = Y => nonnumeric) * visual(Lines) |> draw

This approach is also applicable to DimStacks, since they also convert to DimTables. Let's see an example here.

We'll construct a DimStack with the :data layer being our DimArray A, and an X-only variable :color that we'll use to color the line.

color_vec = DimVector(1:10, X)
ds = DimStack((; data = A, color = color_vec))

data(ds) * mapping(X, :data; color = :color, layout = Y => nonnumeric) * visual(Lines) |> draw

Note

If you wish to subset your DimArray, you can't pass selectors like X(1 .. 2) to AlgebraOfGraphics. Instead, subset the DimArray you pass to data - this is a very cheap operation.

Test series plots

default colormap

B = rand(X(10:10:100), Y([:a, :b, :c, :d, :e, :f, :g, :h, :i, :j]))
Makie.series(B)

A different colormap

The colormap is controlled by the color argument, which can take as an input a named colormap, i.e. :plasma or a list of colours.

Makie.series(B; color=:plasma)

Makie.series(A; color=[:red, :blue, :orange])

with markers

Makie.series(A; color=[:red, :blue, :orange], markersize=15)

A lot more is planned for Makie.jl plots in future!