This is the multi-page printable view of this section. Click here to print.

Return to the regular view of this page.

Cookbooks

Cookbooks for Lisp-Stat

1 - Quick Plot

Building plots with composable, ggplot-style helper functions

Overview

The geom and gg packages provide a set of composable helper functions for building Vega-Lite plot specifications. Inspired by R’s ggplot2, specifications are constructed by combining independent layers — geometry, labels, scales, coordinates and themes — rather than writing monolithic JSON-like plists by hand.

Each helper returns a plist. The function merge-plists recursively merges them into a single Vega-Lite spec, which vega:defplot compiles and plot:plot renders.

Design philosophy

In ggplot2, a plot is built from independent concerns:

ggplot2 Lisp-Stat Package Responsibility
geom_point() point geom Mark type and encoding
geom_bar() bar geom Mark type and encoding
geom_boxplot() box-plot geom Mark type and encoding
geom_histogram() histogram geom Binning, aggregation
geom_line() line geom Series connection, interpolation
labs() label gg Axis titles
scale_*() axes gg Axis transforms, domains, color schemes
coord_cartesian() coord gg Viewport clipping
theme() theme gg Dimensions, fonts, appearance
(no direct equiv.) tooltip gg Hover field definitions

Each function knows about one concern and nothing else. A mark function never sets axis titles; label never touches mark types; theme never alters encodings. This separation means any helper can be used with any plot type.

The merge pattern

Every helper returns a plist fragment. merge-plists performs a recursive deep merge: when two plists both supply a nested plist for the same key (e.g. :encoding), the inner plists are merged rather than one replacing the other. This is what allows label to add :axis entries to the :encoding that point already created.

Note that merge-plists is a utility function, not a layer helper — it is the mechanism that makes composition work, not a layer you pass to qplot yourself. See the API reference for full details.

Two ways to plot

Lisp-Stat provides two entry points for creating plots. Choose the one that fits your workflow.

defplot + plot:plot — the explicit pattern

The traditional approach separates definition from rendering. Use this when you want full control, or when writing scripts and notebooks. Replace :x-field, :y-field, and my-data with your actual field names and data source:

(plot:plot (vega:defplot my-plot (merge-plists `(:title "My Plot" :data (:values ,my-data)) (point :x-field :y-field) (label :x "X Label" :y "Y Label") (theme :width 600))))

defplot is a macro that:

  1. Calls %defplot to create a vega-plot object
  2. Binds it to a global variable (my-plot)
  3. Registers it in *all-plots* so show-plots can list it

plot:plot then renders the object to the browser.

qplot — quick plot for the REPL

For interactive exploration, the three-line scaffold of plot:plot / vega:defplot / merge-plists is repetitive. qplot collapses it into a single function call:

(qplot 'my-plot my-data `(:title "My Plot") (point :x-field :y-field) (label :x "X Label" :y "Y Label") (theme :width 600))

qplot takes a name (a symbol), a data object (a data frame, plist data, or URI), and any number of layer plists. It:

  1. Prepends (:data (:values ,data)) and merges all layers
  2. Creates the plot object via %defplot
  3. Binds it to the named global variable
  4. Registers it in *all-plots*
  5. Renders immediately via plot:plot
  6. Returns the plot object

Because qplot binds a named variable, the standard REPL workflow is to re-evaluate the same form as you iterate on a plot. Each call overwrites the previous definition — there is no accumulation in *all-plots* or in the global namespace:

;; First attempt — rough sketch (qplot 'cars vgcars (point :horsepower :miles-per-gallon)) ;; Second attempt — add color and a title (qplot 'cars vgcars `(:title "HP vs MPG") (point :horsepower :miles-per-gallon :color :origin :filled t)) ;; Third attempt — polish for presentation (qplot 'cars vgcars `(:title "HP vs MPG") (point :horsepower :miles-per-gallon :color :origin :filled t) (label :x "Horsepower" :y "Fuel Efficiency") (theme :width 600 :height 400)) ;; Later — the variable is still bound cars ; => #<VEGA-PLOT ...> (plot:plot cars) ; re-render (describe cars) ; inspect the spec (show-plots) ; lists one 'cars' entry

Helpers reference

The layering helpers work with all mark types.

label — Set axis titles.

(label :x "Horsepower" :y "Miles per Gallon")

axes — Set axis types, domains, ranges, and color schemes.

(axes :x-type :log :color-scheme :dark2) (axes :x-domain #(0 300) :y-domain #(0 50))

tooltip — Add hover tooltips. Each argument is a field spec.

(tooltip '(:field :name :type :nominal) '(:field :horsepower :type :quantitative))

coord — Restrict the visible viewport and clip marks, like coord_cartesian() in ggplot2. Only data within the domain is visible; points outside are clipped rather than dropped.

(coord :x-domain #(50 150) :y-domain #(20 40))

theme — Set dimensions, font, background, or named presets.

(theme :width 600 :height 400 :font "Georgia")

Loading example data

The examples below use datasets from the Vega datasets collection. Load them into your session before running the examples:

(vega:load-vega-examples)

This makes the following variables available in your environment:

  • vgcars — Automobile specifications (horsepower, MPG, origin, etc.) for 406 cars.
  • stocks — Daily closing prices for several major tech stocks over multiple years.

For datasets loaded using vega:read-vega, field names are automatically converted to Lisp-style keywords (e.g. Miles_per_Gallon becomes :miles-per-gallon). Note that the :year field in vgcars is stored as a date string (e.g. "1970-01-01") rather than an integer, which is why the line chart examples pass :x-type :temporal for that field.

Scatter plot

A scatter plot maps two quantitative variables to x and y positions. Use point for exploring relationships, correlations and clusters in data. The name follows the ggplot2 convention where geom_point() produces a scatter plot.

Basic scatter plot

The simplest scatter plot needs just a name, a data frame, and two field names:

(qplot 'cars-basic vgcars `(:title "Horsepower vs. MPG") (point :horsepower :miles-per-gallon))

Color by category

Pass a keyword to :color to map a nominal variable to hue. Use :filled t to fill the point marks:

(qplot 'cars-colored vgcars `(:title "Cars by Origin") (point :horsepower :miles-per-gallon :color :origin :filled t))

Bubble plot

When :size is a keyword, it encodes a third quantitative field as point area — a bubble plot:

(qplot 'cars-bubble vgcars `(:title "Bubble: Size = Acceleration") (point :horsepower :miles-per-gallon :color :origin :size :acceleration :filled t) (label :x "Horsepower" :y "Miles per Gallon"))

Adding axis labels

Use label to give axes meaningful titles:

(qplot 'cars-with-labels vgcars `(:title "Vega Cars") (point :horsepower :miles-per-gallon :filled t) (label :x "Engine Horsepower" :y "Fuel Efficiency (MPG)"))

Log scale with custom color scheme

Use axes to transform an axis and change the color palette:

(qplot 'cars-log-scale vgcars `(:title "Horsepower (log) vs. MPG") (point :horsepower :miles-per-gallon :color :origin :filled t) (axes :x-type :log :color-scheme :dark2) (label :x "Horsepower (log scale)" :y "Miles per Gallon"))

Tooltips on hover

Add tooltip to show details when the user hovers over a point:

(qplot 'cars-tooltip vgcars `(:title "Car Details on Hover") (point :horsepower :miles-per-gallon :color :origin :filled t) (tooltip '(:field :name :type :nominal) '(:field :horsepower :type :quantitative) '(:field :miles-per-gallon :type :quantitative) '(:field :origin :type :nominal)))

Zoom into a region

Use coord to restrict the visible area. Unlike axes, this clips marks that fall outside the domain — only points within the viewport are drawn:

(qplot 'cars-zoomed vgcars `(:title "Cars: 50-150 HP, 20-40 MPG") (point :horsepower :miles-per-gallon :color :origin :filled t) (coord :x-domain #(50 150) :y-domain #(20 40)) (label :x "Horsepower" :y "Miles per Gallon"))

Custom theme

Use theme to set plot dimensions, font, and visual style:

(qplot 'cars-themed vgcars `(:title "Themed Scatter Plot") (point :horsepower :miles-per-gallon :color :origin :filled t) (label :x "Horsepower" :y "MPG") (theme :width 600 :height 400 :font "Georgia"))

Full example — all layers

Combine every layer for a production-quality plot:

(qplot 'cars-full vgcars `(:title "Complete Example: All Layers" :description "Demonstrating label, axes, tooltip, coord, and theme") (point :horsepower :miles-per-gallon :color :origin :size :acceleration :filled t) (label :x "Engine Horsepower" :y "Fuel Efficiency (MPG)") (axes :color-scheme :category10) (tooltip '(:field :name :type :nominal) '(:field :horsepower :type :quantitative) '(:field :miles-per-gallon :type :quantitative) '(:field :origin :type :nominal)) (coord :x-domain #(40 240) :y-domain #(5 50)) (theme :width 700 :height 450))

LOESS smoother

A LOESS (locally estimated scatterplot smoothing) curve fits a non-parametric smooth line through data, revealing trends without assuming a fixed functional form. Use loess to overlay a trend line on a scatter plot or to compare smoothed trajectories across groups.

Scatter plot with LOESS smoother

Use gg:layer to compose the scatter points and the smoother into a single layered view:

(qplot 'cars-loess vgcars `(:title "HP vs. MPG with LOESS Smoother") (gg:layer (point :horsepower :miles-per-gallon :color :origin :filled t :opacity 0.5) (loess :horsepower :miles-per-gallon :group :origin :stroke-width 2)))

The :group :origin argument fits a separate curve for each origin and encodes it with the matching hue automatically. Increase :bandwidth toward 1.0 for a flatter, more global fit; decrease it toward 0.05 for a curve that tracks local variation closely.

Histogram

A histogram bins a quantitative variable and counts observations per bin. Use histogram to visualize the distribution of a single variable.

Basic histogram

Pass a single field name. The default uses Vega-Lite’s automatic binning and counts occurrences:

(qplot 'mpg-hist vgcars `(:title "Distribution of Miles per Gallon") (histogram :miles-per-gallon) (label :x "Miles per Gallon" :y "Count"))

Custom bin count

Control granularity with the :bin keyword. Pass a plist with :maxbins to limit the number of bins:

(qplot 'mpg-hist-bins vgcars `(:title "MPG Distribution (10 bins)") (histogram :miles-per-gallon :bin '(:maxbins 10)) (label :x "Miles per Gallon" :y "Count"))

Horizontal histogram

Set :orient :horizontal to place bins on the y-axis:

(qplot 'mpg-hist-horiz vgcars `(:title "MPG Distribution (Horizontal)") (histogram :miles-per-gallon :orient :horizontal) (label :x "Count" :y "Miles per Gallon"))

Stacked histogram by group

Pass :group to split bins by a nominal field. Vega-Lite automatically stacks the bars:

(qplot 'mpg-hist-stacked vgcars `(:title "MPG Distribution by Origin") (histogram :miles-per-gallon :group :origin) (label :x "Miles per Gallon" :y "Count"))

Layered histogram

Use :stack :null with :opacity to overlay distributions transparently instead of stacking them:

(qplot 'mpg-hist-layered vgcars `(:title "MPG: Overlaid by Origin") (histogram :miles-per-gallon :group :origin :stack :null :opacity 0.5) (label :x "Miles per Gallon" :y "Count"))

Normalized (100%) stacked histogram

Use :stack :normalize to show proportions instead of counts:

(qplot 'mpg-hist-normalized vgcars `(:title "MPG: Proportion by Origin") (histogram :miles-per-gallon :group :origin :stack :normalize) (label :x "Miles per Gallon" :y "Proportion"))

Styled histogram

Use :color, :corner-radius-end, and :bin-spacing for visual polish. Combine with theme for custom dimensions:

(qplot 'mpg-hist-styled vgcars `(:title "Styled Histogram") (histogram :miles-per-gallon :color "darkslategray" :corner-radius-end 3 :bin-spacing 0) (label :x "Miles per Gallon" :y "Count") (theme :width 500 :height 300))

Bar chart

A bar chart maps a categorical variable to position and a quantitative variable to bar length. Use bar when your x-axis is nominal or ordinal rather than a continuous distribution. The name follows the ggplot2 convention where geom_bar() produces a bar chart.

Basic bar chart

Supply the categorical field and the quantitative field:

(qplot 'origin-bar vgcars `(:title "Average MPG by Origin") (bar :origin :miles-per-gallon :aggregate :mean) (label :x "Origin" :y "Mean Miles per Gallon"))

Horizontal bar chart

Set :orient :horizontal to swap axes — useful for long category labels:

(qplot 'origin-bar-horiz vgcars `(:title "Average MPG by Origin (Horizontal)") (bar :origin :miles-per-gallon :aggregate :mean :orient :horizontal) (label :x "Mean Miles per Gallon" :y "Origin"))

Grouped (stacked) bar chart

Pass :group to split bars by a second nominal field. Bars are stacked by default:

(qplot 'cylinders-by-origin vgcars `(:title "Car Count: Cylinders by Origin") (bar :cylinders :miles-per-gallon :aggregate :count :group :origin) (label :x "Cylinders" :y "Count"))

Styled bar chart

Combine visual options with layering helpers:

(qplot 'origin-bar-styled vgcars `(:title "Mean MPG by Origin") (bar :origin :miles-per-gallon :aggregate :mean :color "teal" :corner-radius-end 4) (label :x "Origin" :y "Mean MPG") (theme :width 400 :height 300 :font "Helvetica"))

Box plot

A box plot summarizes the distribution of a quantitative variable, showing the median, interquartile range and outliers. Use box-plot to compare distributions across groups.

1D box plot

A single quantitative field produces a box plot summarizing the entire variable:

(qplot 'mpg-box-1d vgcars `(:title "MPG Distribution") (box-plot :miles-per-gallon) (label :x "Miles per Gallon"))

2D box plot — compare groups

Pass :category to split the box plot by a nominal field:

(qplot 'mpg-box-by-origin vgcars `(:title "MPG by Origin") (box-plot :miles-per-gallon :category :origin) (label :x "Miles per Gallon" :y "Origin"))

Vertical orientation

Set :orient :vertical to place categories on the x-axis and values on the y-axis:

(qplot 'mpg-box-vertical vgcars `(:title "MPG by Cylinders (Vertical)") (box-plot :miles-per-gallon :category :cylinders :orient :vertical) (label :x "Cylinders" :y "Miles per Gallon"))

Min-max whiskers

Set :extent "min-max" to extend whiskers to the minimum and maximum values instead of the default 1.5× IQR (Tukey) whiskers:

(qplot 'mpg-box-minmax vgcars `(:title "MPG by Origin (Min-Max Whiskers)") (box-plot :miles-per-gallon :category :origin :extent "min-max") ; note string value (label :x "Miles per Gallon" :y "Origin"))

Styled box plot

Combine visual options with layering helpers for presentation:

(qplot 'mpg-box-styled vgcars `(:title "MPG by Origin") (box-plot :miles-per-gallon :category :origin :orient :vertical :size 40) (label :x "Origin" :y "Miles per Gallon") (theme :width 500 :height 350))

Line chart

A line chart connects data points in order, typically along a temporal or sequential x-axis. Use line for time series, trends, and any data where the relationship between consecutive points is meaningful. The name follows the ggplot2 convention where geom_line() produces a line chart.

Basic line chart

The simplest line chart needs two field names. Points are connected in x-axis order:

(qplot 'stock-basic stocks `(:title "Google Stock Price" :transform #((:filter "datum.symbol === 'GOOG'"))) (line :date :price :x-type :temporal) (label :x "Date" :y "Price (USD)"))

Multiple series by color

Pass a keyword to :color to draw a separate line for each category:

(qplot 'stock-colored stocks `(:title "Stock Prices by Company") (line :date :price :color :symbol :x-type :temporal) (label :x "Date" :y "Price (USD)"))

Smoothed interpolation

Set :interpolate to control how points are connected. Common values are :linear (default), :monotone (smooth, monotonic curves), :step (step function), :basis (B-spline), and :cardinal:

(qplot 'stock-smooth stocks `(:title "Stock Prices (Smoothed)") (line :date :price :color :symbol :interpolate :monotone :x-type :temporal) (label :x "Date" :y "Price (USD)"))

Line with point markers

Set :point t to overlay point marks on each data position — useful for sparse data or when exact values matter. Note that :x-type :temporal is required here because the :year field in vgcars is stored as a date string rather than an integer:

(qplot 'mpg-trend vgcars `(:title "Mean MPG by Model Year") (line :year :miles-per-gallon :point t :x-type :temporal) (label :x "Model Year" :y "Miles per Gallon"))

Custom stroke width

Use :stroke-width to set a fixed line thickness:

(qplot 'stock-thick stocks `(:title "AAPL Stock Price" :transform #((:filter "datum.symbol === 'AAPL'"))) (line :date :price :stroke-width 3 :x-type :temporal) (label :x "Date" :y "Price (USD)") (theme :width 600 :height 300))

Dashed lines

Use :stroke-dash with a vector to create dashed or dotted lines. The vector specifies alternating dash and gap lengths:

(qplot 'stock-dashed stocks `(:title "Stock Prices (Dashed)") (line :date :price :color :symbol :stroke-dash #(6 3) :opacity 0.8 :x-type :temporal) (label :x "Date" :y "Price (USD)"))

Step chart

Use :interpolate :step for piecewise-constant lines — useful for data that changes at discrete intervals (e.g. interest rates, pricing tiers):

(qplot 'mpg-step vgcars `(:title "MPG by Year (Step)") (line :year :miles-per-gallon :interpolate :step :x-type :temporal) (label :x "Model Year" :y "Miles per Gallon"))

Styled multi-series line chart

Combine all options with layering helpers for a polished presentation:

(qplot 'stock-full stocks `(:title "Stock Comparison" :description "Daily closing prices for major tech stocks") (line :date :price :color :symbol :interpolate :monotone :stroke-width 2 :x-type :temporal) (label :x "Date" :y "Closing Price (USD)") (axes :color-scheme :dark2) (tooltip '(:field :symbol :type :nominal) '(:field :date :type :temporal) '(:field :price :type :quantitative)) (theme :width 700 :height 400))

Function curves

geom:func plots a Lisp function as a smooth line by evaluating it at evenly-spaced sample points and embedding the resulting (x, y) pairs directly in the Vega-Lite specification. It mirrors the behaviour of R’s geom_function() from ggplot2.

Unlike the data-driven helpers (point, bar, histogram, etc.), func is self-contained: it carries its own :data block and requires no external data frame. Use it anywhere you want to visualise a mathematical relationship — probability densities, regression curves, physical models, or any other computable function.

Import func alongside the other helpers you use:

(import '(geom:func))

How it works

func calls aops:linspace to generate n evenly-spaced x values over the closed interval [xmin, xmax] specified by :xlim. It then calls fn at each x, collects the (x, y) pairs into a vector of plists, and embeds them as an inline Vega-Lite :data block. A :line mark connects the points using the chosen interpolation method (:monotone by default, giving smooth curves without overshoot).

Points where fn signals a condition (e.g. (log 0), (/ 1 0)) or returns a non-finite value (± infinity, NaN) are silently dropped. Vega-Lite renders a visible gap at each discontinuity — the correct visual for functions like tan or 1/x.

Design note: self-contained data

All other geom helpers return only :mark and :encoding keys and rely on the caller to supply :data. func also returns a :data key, because the data is the function. This means it composes slightly differently from the other geom helpers:

Helper Data source Typical entry point
point, bar, histogram, … external data frame qplot
func self-generated inline defplot + vega:merge-plists

For multi-layer plots (function overlaid on data) use Vega-Lite’s :layer array directly inside defplot; see Overlay on scatter data below.

The following table describes the ggplot2 equivalent and responsibility:

ggplot2 Lisp-Stat Package Responsibility
geom_function() func geom Sample a function, encode as a line

Reference

(geom:func fn &key xlim n color stroke-width stroke-dash opacity interpolate)
Parameter Type Default Description
fn function A Lisp function (real → real). Receives a double-float; must return a real.
:xlim vector #(0d0 1d0) Domain #(xmin xmax). Both endpoints are always sampled.
:n integer ≥ 2 100 Number of sample points. Increase for oscillatory functions.
:color string nil CSS color for the line, e.g. "steelblue" or "#e63946". nil lets Vega-Lite choose.
:stroke-width number nil Line thickness in pixels.
:stroke-dash vector nil Dash/gap pattern, e.g. #(6 3) for dashes or #(2 2) for dots.
:opacity number 0–1 nil Line opacity.
:interpolate keyword :monotone Vega-Lite interpolation method. :linear, :basis, :cardinal, :step are also accepted.

Basic function plot

Supply a function and a domain. vega:merge-plists combines the self-contained func layer with a title and axis labels:

(vega:defplot sine-wave (vega:merge-plists `(:title "Sine Wave") (func #'sin :xlim #(-6.283 6.283) :n 200) (label :x "x" :y "sin(x)")))

Custom domain and resolution

Increase :n for functions that oscillate rapidly. Use :xlim to set the evaluation domain precisely:

(vega:defplot damped-oscillation (vega:merge-plists `(:title "Damped Oscillation") (func (lambda (x) (* (exp (* -0.3 x)) (sin (* 4 x)))) :xlim #(0 20) :n 400) (label :x "t" :y "Amplitude")))

Functions with singularities

Points where fn raises a condition or returns ±infinity are silently dropped. Vega-Lite draws a gap at each discontinuity — the correct rendering for functions like tan(x):

(vega:defplot tangent-curve (vega:merge-plists `(:title "tan(x) — gaps at singularities") (func #'tan :xlim #(-4.5 4.5) :n 500) (label :x "x" :y "tan(x)") (axes :y-domain #(-10 10))))

Probability density function

Plot a probability density function using the distributions system. Load it before running these examples:

(asdf:load-system :distributions)

Create a distribution object with distributions:r-normal, then pass its pdf method to func. r-normal takes mean and variance (not standard deviation), so the standard normal is (r-normal 0d0 1d0):

(let ((d (distributions:r-normal 0d0 1d0))) ; mean=0, variance=1 (vega:defplot normal-pdf (vega:merge-plists `(:title "Standard Normal PDF") (func (lambda (x) (distributions:pdf d x)) :xlim #(-4 4) :n 200) (label :x "x" :y "Density") (theme :width 500 :height 300))))

Styled function curve

Pass :color, :stroke-width, and :stroke-dash for visual polish. To show two styled curves together, use :layer — each func call carries its own inline data:

(vega:defplot sin-and-cos-styled `(:title "sin and cos — styled lines" :data (:values #()) :layer #(,(func #'sin :xlim #(-6.283 6.283) :n 200 :color "steelblue" :stroke-width 2) ,(func #'cos :xlim #(-6.283 6.283) :n 200 :color "firebrick" :stroke-width 2 :stroke-dash #(8 4)))))

Step interpolation

Use :interpolate :step for piecewise-constant functions — useful for visualising floor, ceiling, or any discrete-valued function:

(vega:defplot step-function (vega:merge-plists `(:title "Floor function") (func #'ffloor :xlim #(0 6) :n 300 :interpolate :step :color "darkslategray" :stroke-width 2) (label :x "x" :y "floor(x)")))

Two functions on the same axes

Use Vega-Lite’s :layer array directly inside defplot. Each func call produces an independent layer with its own inline data:

(vega:defplot sin-vs-cos `(:title "sin(x) and cos(x)" :data (:values #()) :layer #(;; Layer 1 — sine ,(func #'sin :xlim #(-6.283 6.283) :n 200 :color "steelblue" :stroke-width 2) ;; Layer 2 — cosine ,(func #'cos :xlim #(-6.283 6.283) :n 200 :color "firebrick" :stroke-width 2 :stroke-dash #(6 3)))))

Family of curves

Plot a parameterised family by building the :layer vector in a loop:

;; Gaussian PDFs with increasing standard deviations (let* ((sigmas #(0.5 1.0 1.5 2.0)) (colors #("steelblue" "seagreen" "darkorange" "firebrick")) (layers (map 'vector (lambda (sigma color) (func (lambda (x) (* (/ 1 (* sigma (sqrt (* 2 pi)))) (exp (* -0.5 (expt (/ x sigma) 2))))) :xlim #(-5 5) :n 200 :color color)) sigmas colors))) (vega:defplot gaussian-family `(:title "Gaussian PDFs for sigma = 0.5, 1, 1.5, 2" :data (:values #()) :layer ,layers)))

Overlay on scatter data

When overlaying a function on top of data, each layer supplies its own :data. The data layer uses the original data frame; the function layer uses the inline data generated by func:

(vega:defplot cars-with-trend `(:title "HP vs MPG with Quadratic Trend" :data (:values #()) :layer #(;; Layer 1: raw data as a scatter plot (:data (:values ,vgcars) :mark (:type :point :filled t :opacity 0.5) :encoding (:x (:field :horsepower :type :quantitative :title "Horsepower") :y (:field :miles-per-gallon :type :quantitative :title "Miles per Gallon") :color (:field :origin :type :nominal))) ;; Layer 2: fitted quadratic y = 52 - 0.23x + 3e-4*x^2 ,(func (lambda (x) (+ 52.0d0 (* -0.23d0 x) (* 3.0d-4 (expt x 2)))) :xlim #(40 230) :n 300 :color "firebrick" :stroke-width 2.5))))

Normal distribution fit over a histogram

Overlay the theoretical PDF on an empirical histogram. Use select:select to extract the column as a vector, statistics:sd for the standard deviation, and distributions:r-normal to construct the fitted distribution — recall that r-normal takes the variance, so pass (expt sigma 2):

(let* ((mpg (select:select vgcars t :miles-per-gallon)) (mu (statistics:mean mpg)) (sigma (statistics:sd mpg)) (d (distributions:r-normal mu (expt sigma 2)))) (vega:defplot mpg-fit `(:title "MPG: Empirical Histogram with Normal Fit" :data (:values #()) :layer #(;; Histogram layer (:data (:values ,vgcars) :mark :bar :encoding (:x (:field :miles-per-gallon :bin (:maxbins 15) :type :quantitative :title "Miles per Gallon") :y (:aggregate :count :stack :null :type :quantitative :title "Count"))) ;; Density curve — scaled by (n × bin-width) to match count axis ,(func (lambda (x) (* 406 3 (distributions:pdf d x))) :xlim #(5 50) :n 300 :color "firebrick" :stroke-width 2)))))

Chebyshev approximation

numerical-utilities provides chebyshev-regression and evaluate-chebyshev for polynomial approximation. Plot the exact function and its approximation together to inspect accuracy:

(let* ((coeffs (nu:chebyshev-regression (lambda (x) (exp (- (nu:square x)))) 12)) ; 12-term approximation (approx (lambda (x) (nu:evaluate-chebyshev coeffs x)))) (vega:defplot chebyshev-approx `(:title "exp(-x^2): Exact vs 12-Term Chebyshev Approximation" :data (:values #()) :layer #(,(func (lambda (x) (exp (- (nu:square x)))) :xlim #(-1 1) :n 200 :color "steelblue" :stroke-width 2) ,(func approx :xlim #(-1 1) :n 200 :color "firebrick" :stroke-width 1.5 :stroke-dash #(6 3))))))

Combining layers across plot types

Because every helper returns an independent plist, you can mix and match freely. Here are patterns that work with all mark types.

Labels on any plot

label works identically on every plot type — it only touches :encoding :x/:y :axis :title. The examples below produce the same visual results as the per-section examples above; they are shown here to illustrate that label requires no changes across mark types:

;; On a scatter plot (qplot 'point-labeled vgcars (point :horsepower :miles-per-gallon) (label :x "Engine Horsepower" :y "Fuel Efficiency"))
;; On a histogram (qplot 'hist-labeled vgcars (histogram :horsepower) (label :x "Engine Horsepower" :y "Number of Cars"))
;; On a bar chart (qplot 'bar-labeled vgcars (bar :origin :miles-per-gallon :aggregate :mean) (label :x "Country of Origin" :y "Average MPG"))
;; On a box plot (qplot 'box-labeled vgcars (box-plot :miles-per-gallon :category :origin) (label :x "Fuel Efficiency" :y "Manufacturing Origin"))

Theme on any plot

theme sets top-level properties that apply to every plot type:

(qplot 'hist-themed vgcars `(:title "Horsepower Distribution") (histogram :horsepower :color "darkslategray") (label :x "Horsepower" :y "Count") (theme :width 500 :height 300 :font "Georgia"))

coord with continuous axes

coord works best with quantitative (continuous) axes. For bar charts and box plots whose axes are categorical, use coord with :clip nil to avoid cutting marks, or use axes with :x-domain or :y-domain instead:

;; Zoom a scatter plot — clip marks outside the viewport (coord :x-domain #(50 150) :y-domain #(20 40)) ;; Restrict a bar chart's quantitative axis — no clipping (coord :y-domain #(0 35) :clip nil) ;; Alternatively, use axes for categorical axes (axes :y-domain #(0 35))

Recipe: exploring a new dataset

When you first load a dataset, a quick exploratory sequence might look like this. Note how the same plot name can be reused — each qplot call overwrites the previous definition.

Step 1 — Distribution of a key variable

Start with a histogram to understand the shape of the data:

(qplot 'explore-1 vgcars `(:title "Horsepower Distribution") (histogram :horsepower))

Step 2 — Scatter plot of two main variables

Look for relationships between variables:

(qplot 'explore-2 vgcars `(:title "HP vs. MPG") (point :horsepower :miles-per-gallon :filled t))

Step 3 — Compare groups

Break the data down by category to see if patterns differ:

(qplot 'explore-3 vgcars `(:title "MPG by Origin") (box-plot :miles-per-gallon :category :origin :orient :vertical) (label :x "Origin" :y "MPG"))

Step 4 — Add detail

Color by group and add labels to confirm the story:

(qplot 'explore-4 vgcars `(:title "HP vs. MPG by Origin") (point :horsepower :miles-per-gallon :color :origin :filled t) (label :x "Horsepower" :y "Miles per Gallon") (tooltip '(:field :name :type :nominal) '(:field :origin :type :nominal) '(:field :horsepower :type :quantitative) '(:field :miles-per-gallon :type :quantitative)))

At the REPL, every call overwrites the same explore variable. The variable always holds the latest version:

explore ; => #<VEGA-PLOT ...> (show-plots) ; one entry for EXPLORE

Recipe: presentation-ready plot

For reports or publications, combine all the layering helpers:

(qplot 'presentation vgcars `(:title "Automobile Performance by Country of Origin" :description "Horsepower vs fuel efficiency for 406 cars") (point :horsepower :miles-per-gallon :color :origin :filled t) (label :x "Engine Horsepower" :y "Fuel Efficiency (MPG)") (axes :color-scheme :tableau10) (tooltip '(:field :name :type :nominal) '(:field :horsepower :type :quantitative) '(:field :miles-per-gallon :type :quantitative)) (theme :width 700 :height 450 :font "Helvetica"))

Recipe: saving a plot for later

Once you are happy with a plot created via qplot, it is a first-class vega-plot object. You can re-render, inspect, or write it to a file. This example assumes presentation was created in the recipe above:

;; Re-render in the browser (plot:plot presentation) ;; Inspect the spec (describe presentation) ;; List all named plots (show-plots)

Recipe: exploring a function

A typical REPL workflow when investigating a mathematical function. Note how the same plot name is reused — each defplot call overwrites the previous definition.

Step 1 — Quick sketch over the natural domain

(vega:defplot explore-fn (vega:merge-plists `(:title "First look") (func (lambda (x) (/ (sin x) x)) ; sinc :xlim #(-20 20) :n 400)))

Step 2 — Zoom in on a region of interest

(vega:defplot explore-fn (vega:merge-plists `(:title "sinc(x) — detail near origin") (func (lambda (x) (/ (sin x) x)) :xlim #(-6.283 6.283) :n 400) (label :x "x" :y "sin(x)/x") (axes :y-domain #(-0.3 1.1))))

Step 3 — Presentation polish

(vega:defplot sinc-final (vega:merge-plists `(:title "sinc(x) = sin(x)/x" :description "Classic sinc function over [-2pi, 2pi]") (func (lambda (x) (/ (sin x) x)) :xlim #(-6.283 6.283) :n 500 :color "steelblue" :stroke-width 2) (label :x "x" :y "sin(x) / x") (theme :width 600 :height 350 :font "Georgia")))

Recipe: comparing model fits

Overlay multiple fitted curves on a scatter plot to compare competing models visually:

(let* ((linear (lambda (x) (+ 39.9 (* -0.158 x)))) (quadratic (lambda (x) (+ 52.0 (* -0.23 x) (* 3.0e-4 (expt x 2)))))) (vega:defplot model-comparison `(:title "Linear vs Quadratic Fit" :data (:values #()) :layer #(;; Raw data (:data (:values ,vgcars) :mark (:type :point :filled t :color "lightgray") :encoding (:x (:field :horsepower :type :quantitative :title "Horsepower") :y (:field :miles-per-gallon :type :quantitative :title "Miles per Gallon"))) ;; Linear fit ,(func linear :xlim #(40 230) :n 200 :color "steelblue" :stroke-width 2) ;; Quadratic fit ,(func quadratic :xlim #(40 230) :n 200 :color "firebrick" :stroke-width 2 :stroke-dash #(6 3))))))

Recipe: dropping down to raw Vega-Lite

The helpers cover common patterns. When you need something they don’t support — transforms, selections, parameters, calculated fields, multi-view compositions — you can write raw Vega-Lite directly.

Example: brush selection with linked data table

This example reproduces the Vega-Lite brush table: drag a rectangle over the scatter plot to see the selected cars in a table. It uses hconcat (side-by-side views), params (interactive brush), and transform (filter + rank) — none of which the helpers generate. For specs this complex, write the full Vega-Lite plist and use defplot directly:

(plot:plot (vega:defplot brush-table `(:description "Drag a rectangular brush to show selected points in a table." :data (:values ,vgcars) :transform #((:window #((:op :row-number :as "row_number")))) :hconcat #(;; Left panel: scatter plot with brush (:params #((:name "brush" :select "interval")) :mark :point :encoding (:x (:field :horsepower :type :quantitative) :y (:field :miles-per-gallon :type :quantitative) :color (:condition (:param "brush" :field :cylinders :type :ordinal) :value "grey"))) ;; Right panel: table of selected points (:transform #((:filter (:param "brush")) (:window #((:op :rank :as "rank"))) (:filter (:field "rank" :lt 20))) :hconcat #((:width 50 :title "Horsepower" :mark :text :encoding (:text (:field :horsepower :type :nominal) :y (:field "row_number" :type :ordinal :axis :null))) (:width 50 :title "MPG" :mark :text :encoding (:text (:field :miles-per-gallon :type :nominal) :y (:field "row_number" :type :ordinal :axis :null))) (:width 50 :title "Origin" :mark :text :encoding (:text (:field :origin :type :nominal) :y (:field "row_number" :type :ordinal :axis :null)))))) :resolve (:legend (:color :independent)))))

This plot cannot be built with qplot and the layering helpers; the hconcat layout requires two independent views with different marks, encodings, and transforms. The rule of thumb:

  • Use qplot + helpers for single-view plots (scatter, line, bar, histogram, box plot) with standard encodings
  • Use defplot + raw plists for multi-view layouts (hconcat, vconcat, layer, facet), complex interactions, or any Vega-Lite feature the helpers don’t cover

Both approaches produce the same vega-plot objects and coexist in the same session.

2 - Vega Plotting

Example plots using the vega-lite DSL.

The plots here show equivalents to the Vega-Lite example gallery. Before you begin working with these example, be certain to read the plotting tutorial where you will learn the basics of working with plot specifications and data.

Preliminaries

Load Vega-Lite

Load Vega-Lite and network libraries:

(asdf:load-system :plot/vega)

and change to the Lisp-Stat user package:

(in-package :ls-user)

Load example data

The examples in this section use the vega-lite data sets. Load them all now:

(vega:load-vega-examples)

Bar charts

Bar charts are used to display information about categorical variables.

Simple bar chart

In this simple bar chart example we’ll demonstrate using literal embedded data in the form of a plist. Later you’ll see how to use a data-frame directly.

(plot:plot (vega:defplot simple-bar-chart `(:mark :bar :data (:values ,(plist-df '(:a #(A B C D E F G H I) :b #(28 55 43 91 81 53 19 87 52)))) :encoding (:x (:field :a :type :nominal :axis ("labelAngle" 0)) :y (:field :b :type :quantitative)))))

Grouped bar chart

(plot:plot (vega:defplot grouped-bar-chart `(:mark :bar :data (:values ,(plist-df '(:category #(A A A B B B C C C) :group #(x y z x y z x y z) :value #(0.1 0.6 0.9 0.7 0.2 1.1 0.6 0.1 0.2)))) :encoding (:x (:field :category) :y (:field :value :type :quantitative) :x-offset (:field :group) :color (:field group)))))

Stacked bar chart

This example uses Seattle weather from the Vega website. Load it into a data frame like so:

(defdf seattle-weather (read-csv vega:seattle-weather)) ;=> #<DATA-FRAME (1461 observations of 6 variables)>

We’ll use a data-frame as the data source via the Common Lisp backquote mechanism. The spec list begins with a backquote (`) and then the data frame is inserted as a literal value with a comma (,). We’ll use this pattern frequently.

(plot:plot (vega:defplot stacked-bar-chart `(:mark :bar :data (:values ,seattle-weather) :encoding (:x (:time-unit :month :field :date :type :ordinal :title "Month of the year") :y (:aggregate :count :type :quantitative) :color (:field :weather :type :nominal :title "Weather type" :scale (:domain #("sun" "fog" "drizzle" "rain" "snow") :range #("#e7ba52", "#c7c7c7", "#aec7e8", "#1f77b4", "#9467bd")))))))

Population pyramid

Vega calls this a diverging stacked bar chart. It is a population pyramid for the US in 2000, created using the stack feature of vega-lite. You could also create one using concat.

First, load the population data if you haven’t done so:

(defdf population (vega:read-vega vega:population)) ;=> #<DATA-FRAME (570 observations of 4 variables)>

Note the use of read-vega in this case. This is because the data in the Vega example is in an application specific JSON format (Vega, of course).

(plot:plot (vega:defplot pyramid-bar-chart `(:mark :bar :data (:values ,population) :width 300 :height 200 :transform #((:filter "datum.year == 2000") (:calculate "datum.sex == 2 ? 'Female' : 'Male'" :as :gender) (:calculate "datum.sex == 2 ? -datum.people : datum.people" :as :signed-people)) :encoding (:x (:aggregate :sum :field :signed-people :title "population") :y (:field :age :axis nil :sort :descending) :color (:field :gender :scale (:range #("#675193" "#ca8861")))) :config (:view (:stroke nil) :axis (:grid :false)))))

Histograms & density

Basic

For this simple histogram example we’ll use the IMDB film rating data set.

(plot:plot (vega:defplot imdb-plot `(:mark :bar :data (:values ,imdb) :encoding (:x (:bin (:maxbins 8) :field :imdb-rating) :y (:aggregate :count)))))

Relative frequency

Use a relative frequency histogram to compare data sets with different numbers of observations.

The data is binned with first transform. The number of values per bin and the total number are calculated in the second and the third transform to calculate the relative frequency in the last transformation step.

(plot:plot (vega:defplot relative-frequency-histogram `(:title "Relative Frequency" :data (:values ,vgcars) :transform #((:bin t :field :horsepower :as #(:bin-horsepower :bin-horsepower-end)) (:aggregate #((:op :count :as "Count")) :groupby #(:bin-horsepower :bin-horsepower-end)) (:joinaggregate #((:op :sum :field "Count" :as "TotalCount"))) (:calculate "datum.Count/datum.TotalCount" :as :percent-of-total)) :mark (:type :bar :tooltip t) :encoding (:x (:field :bin-horsepower :title "Horsepower" :bin (:binned t)) :x2 (:field :bin-horsepower-end) :y (:field :percent-of-total :type "quantitative" :title "Relative Frequency" :axis (:format ".1~%"))))))

2D histogram scatterplot

If you haven’t already loaded the imdb data set, do so now:

(defparameter imdb (vega:read-vega vega:movies))
(plot:plot (vega:defplot histogram-scatterplot `(:mark :circle :data (:values ,imdb) :encoding (:x (:bin (:maxbins 10) :field :imdb-rating) :y (:bin (:maxbins 10) :field :rotten-tomatoes-rating) :size (:aggregate :count)))))

Stacked density

(plot:plot (vega:defplot stacked-density `(:title "Distribution of Body Mass of Penguins" :width 400 :height 80 :data (:values ,penguins) :mark :bar :transform #((:density |BODY-MASS-(G)| :groupby #(:species) :extent #(2500 6500))) :encoding (:x (:field :value :type :quantitative :title "Body Mass (g)") :y (:field :density :type :quantitative :stack :zero) :color (:field :species :type :nominal)))))

Note the use of the multiple escape characters (|) surrounding the field BODY-MASS-(G). This is required because the JSON data set has parenthesis in the variable names, and these are reserved characters in Common Lisp. The JSON importer wrapped these in the escape character.

Scatter plots

Basic

A basic Vega-Lite scatterplot showing horsepower and miles per gallon for various cars.

(plot:plot (vega:defplot hp-mpg `(:title "Horsepower vs. MPG" :data (:values ,vgcars) :mark :point :encoding (:x (:field :horsepower :type "quantitative") :y (:field :miles-per-gallon :type "quantitative")))))

Colored

In this example we’ll show how to add additional information to the cars scatter plot to show the cars origin. The Vega-Lite example shows that we have to add two new directives to the encoding of the plot:

(plot:plot (vega:defplot hp-mpg-plot `(:title "Vega Cars" :data (:values ,vgcars) :mark :point :encoding (:x (:field :horsepower :type "quantitative") :y (:field :miles-per-gallon :type "quantitative") :color (:field :origin :type "nominal") :shape (:field :origin :type "nominal")))))

With this change we can see that the higher horsepower, lower efficiency, cars are from the USA, and the higher efficiency cars from Japan and Europe.

Text marks

The same information, but further indicated with a text marker. This Vega-Lite example uses a data transformation.

(plot:plot (vega:defplot colored-text-hp-mpg-plot `(:title "Vega Cars" :data (:values ,vgcars) :transform #((:calculate "datum.origin[0]" :as "OriginInitial")) :mark :text :encoding (:x (:field :horsepower :type "quantitative") :y (:field :miles-per-gallon :type "quantitative") :color (:field :origin :type "nominal") :text (:field "OriginInitial" :type "nominal")))))

Notice here we use a string for the field value and not a symbol. This is because Vega is case sensitive, whereas Lisp is not. We could have also used a lower-case :as value, but did not to highlight this requirement for certain Vega specifications.

Mean & SD overlay

This Vega-Lite scatterplot with mean and standard deviation overlay demonstrates the use of layers in a plot.

Lisp-Stat equivalent

(plot:plot (vega:defplot mean-hp-mpg-plot `(:title "Vega Cars" :data (:values ,vgcars) :layer #((:mark :point :encoding (:x (:field :horsepower :type "quantitative") :y (:field :miles-per-gallon :type "quantitative"))) (:mark (:type :errorband :extent :stdev :opacity 0.2) :encoding (:y (:field :miles-per-gallon :type "quantitative" :title "Miles per Gallon"))) (:mark :rule :encoding (:y (:field :miles-per-gallon :type "quantitative" :aggregate :mean)))))))

Linear regression

(plot:plot (vega:defplot linear-regression `(:data (:values ,imdb) :layer #((:mark (:type :point :filled t) :encoding (:x (:field :rotten-tomatoes-rating :type :quantitative :title "Rotten Tomatoes Rating") :y (:field :imdb-rating :type :quantitative :title "IMDB Rating"))) (:mark (:type :line :color "firebrick") :transform #((:regression :imdb-rating :on :rotten-tomatoes-rating)) :encoding (:x (:field :rotten-tomatoes-rating :type :quantitative :title "Rotten Tomatoes Rating") :y (:field :imdb-rating :type :quantitative :title "IMDB Rating"))) (:transform #((:regression :imdb-rating :on :rotten-tomatoes-rating :params t) (:calculate "'R²: '+format(datum.rSquared, '.2f')" :as :r2)) :mark (:type :text :color "firebrick" :x :width :align :right :y -5) :encoding (:text (:type :nominal :field :r2)))))))

Loess regression

(plot:plot (vega:defplot loess-regression `(:data (:values ,imdb) :layer #((:mark (:type :point :filled t) :encoding (:x (:field :rotten-tomatoes-rating :type :quantitative :title "Rotten Tomatoes Rating") :y (:field :imdb-rating :type :quantitative :title "IMDB Rating"))) (:mark (:type :line :color "firebrick") :transform #((:loess :imdb-rating :on :rotten-tomatoes-rating)) :encoding (:x (:field :rotten-tomatoes-rating :type :quantitative :title "Rotten Tomatoes Rating") :y (:field :imdb-rating :type :quantitative :title "IMDB Rating")))))))

Residuals

A dot plot showing each film in the database, and the difference from the average movie rating. The display is sorted by year to visualize everything in sequential order. The graph is for all films before 2019. Note the use of the filter-rows function.

(plot:plot (vega:defplot residuals `(:data (:values ,(filter-rows imdb '(and (not (eql imdb-rating :na)) (local-time:timestamp< release-date (local-time:parse-timestring "2019-01-01"))))) :transform #((:joinaggregate #((:op :mean :field :imdb-rating :as :average-rating))) (:calculate "datum['imdbRating'] - datum.averageRating" :as :rating-delta)) :mark :point :encoding (:x (:field :release-date :type :temporal :title "Release Date") :y (:field :rating-delta :type :quantitative :title "Rating Delta") :color (:field :rating-delta :type :quantitative :scale (:domain-mid 0) :title "Rating Delta")))))

Query

The cars scatterplot allows you to see miles per gallon vs. horsepower. By adding sliders, you can select points by the number of cylinders and year as well, effectively examining 4 dimensions of data. Drag the sliders to highlight different points.

(plot:plot (vega:defplot scatter-queries `(:data (:values ,vgcars) :transform #((:calculate "year(datum.year)" :as :year)) :layer #((:params #((:name :cyl-year :value #((:cylinders 4 :year 1799)) :select (:type :point :fields #(:cylinders :year)) :bind (:cylinders (:input :range :min 3 :max 8 :step 1) :year (:input :range :min 1969 :max 1981 :step 1)))) :mark :circle :encoding (:x (:field :horsepower :type :quantitative) :y (:field :miles-per-gallon :type :quantitative) :color (:condition (:param :cyl-year :field :origin :type :nominal) :value "grey"))) (:transform #((:filter (:param :cyl-year))) :mark :circle :encoding (:x (:field :horsepower :type :quantitative) :y (:field :miles-per-gallon :type :quantitative) :color (:field :origin :type :nominal) :size (:value 100)))))))

You can add external links to plots.

(plot:plot (vega:defplot scatter-external-links `(:data (:values ,vgcars) :mark :point :transform #((:calculate "'https://www.google.com/search?q=' + datum.name", :as :url)) :encoding (:x (:field :horsepower :type :quantitative) :y (:field :miles-per-gallon :type :quantitative) :color (:field :origin :type :nominal) :tooltip (:field :name :type :nominal) :href (:field :url :type :nominal)))))

Strip plot

The Vega-Lite strip plot example shows the relationship between horsepower and the number of cylinders using tick marks.

(plot:plot (vega:defplot strip-plot `(:title "Vega Cars" :data (:values ,vgcars) :mark :tick :encoding (:x (:field :horsepower :type :quantitative) :y (:field :cylinders :type :ordinal)))))

1D strip plot

(plot:plot (vega:defplot 1d-strip-plot `(:title "Seattle Precipitation" :data (:values ,seattle-weather) :mark :tick :encoding (:x (:field :precipitation :type :quantitative)))))

Bubble plot

This Vega-Lite example is a visualization of global deaths from natural disasters. A copy of the chart from Our World in Data.

(plot:plot (vega:defplot natural-disaster-deaths `(:title "Deaths from global natural disasters" :width 600 :height 400 :data (:values ,(filter-rows disasters '(not (string= entity "All natural disasters")))) :mark (:type :circle :opacity 0.8 :stroke :black :stroke-width 1) :encoding (:x (:field :year :type :temporal :axis (:grid :false)) :y (:field :entity :type :nominal :axis (:title "")) :size (:field :deaths :type :quantitative :title "Annual Global Deaths" :legend (:clip-height 30) :scale (:range-max 5000)) :color (:field :entity :type :nominal :legend nil)))))

Note how we modified the example by using a lower case entity in the filter to match our default lower case variable names. Also note how we are explicit with parsing the year field as a temporal column. This is because, when creating a chart with inline data, Vega-Lite will parse the field as an integer instead of a date.

Line plots

Simple

(plot:plot (vega:defplot simple-line-plot `(:title "Google's stock price from 2004 to early 2010" :data (:values ,(filter-rows stocks '(string= symbol "GOOG"))) :mark :line :encoding (:x (:field :date :type :temporal) :y (:field :price :type :quantitative)))))

Point markers

By setting the point property of the line mark definition to an object defining a property of the overlaying point marks, we can overlay point markers on top of line.

(plot:plot (vega:defplot point-mark-line-plot `(:title "Stock prices of 5 Tech Companies over Time" :data (:values ,stocks) :mark (:type :line :point t) :encoding (:x (:field :date :time-unit :year) :y (:field :price :type :quantitative :aggregate :mean) :color (:field :symbol :type nominal)))))

Multi-series

This example uses the custom symbol encoding for variables to generate the proper types and labels for x, y and color channels.

(plot:plot (vega:defplot multi-series-line-chart `(:title "Stock prices of 5 Tech Companies over Time" :data (:values ,stocks) :mark :line :encoding (:x (:field stocks:date) :y (:field stocks:price) :color (:field stocks:symbol)))))

Step

(plot:plot (vega:defplot step-chart `(:title "Google's stock price from 2004 to early 2010" :data (:values ,(filter-rows stocks '(string= symbol "GOOG"))) :mark (:type :line :interpolate "step-after") :encoding (:x (:field stocks:date) :y (:field stocks:price)))))

Stroke-dash

(plot:plot (vega:defplot stroke-dash `(:title "Stock prices of 5 Tech Companies over Time" :data (:values ,stocks) :mark :line :encoding (:x (:field stocks:date) :y (:field stocks:price) :stroke-dash (:field stocks:symbol)))))

Confidence interval

Line chart with a confidence interval band.

(plot:plot (vega:defplot line-chart-ci `(:data (:values ,vgcars) :encoding (:x (:field :year :time-unit :year)) :layer #((:mark (:type :errorband :extent :ci) :encoding (:y (:field :miles-per-gallon :type :quantitative :title "Mean of Miles per Gallon (95% CIs)"))) (:mark :line :encoding (:y (:field :miles-per-gallon :aggregate :mean)))))))

Area charts

Simple

(plot:plot (vega:defplot area-chart `(:title "Unemployment across industries" :width 300 :height 200 :data (:values ,unemployment-ind) :mark :area :encoding (:x (:field :date :time-unit :yearmonth :axis (:format "%Y")) :y (:field :count :aggregate :sum :title "count")))))

Stacked

Stacked area plots

(plot:plot (vega:defplot stacked-area-chart `(:title "Unemployment across industries" :width 300 :height 200 :data (:values ,unemployment-ind) :mark :area :encoding (:x (:field :date :time-unit :yearmonth :axis (:format "%Y")) :y (:field :count :aggregate :sum :title "count") :color (:field :series :scale (:scheme "category20b"))))))

Horizon graph

A horizon graph is a technique for visualising time series data in a manner that makes comparisons easier. It is based on work done at the UW Interactive Data Lab. See Sizing the Horizon: The Effects of Chart Size and Layering on the Graphical Perception of Time Series Visualizations for more details on Horizon Graphs.

(plot:plot (vega:defplot horizon-graph `(:title "Horizon graph with 2 layers" :width 300 :height 50 :data (:values ,(plist-df `(:x ,(aops:linspace 1 20 20) :y #(28 55 43 91 81 53 19 87 52 48 24 49 87 66 17 27 68 16 49 15)))) :encoding (:x (:field :x :scale (:zero :false :nice :false)) :y (:field :y :type :quantitative :scale (:domain #(0 50)) :axis (:title "y"))) :layer #((:mark (:type :area :clip t :orient :vertical :opacity 0.6)) (:transform #((:calculate "datum.y - 50" :as :ny)) :mark (:type :area :clip t :orient :vertical) :encoding (:y (:field "ny" :type :quantitative :scale (:domain #(0 50))) :opacity (:value 0.3)))) :config (:area (:interpolate :monotone)))))

With overlay

Area chart with overlaying lines and point markers.

(plot:plot (vega:defplot area-with-overlay `(:title "Google's stock price" :data (:values ,(filter-rows stocks '(string= symbol "GOOG"))) :mark (:type :area :line t :point t) :encoding (:x (:field stocks:date) :y (:field stocks:price)))))

Note the use of the variable symbols, e.g. stocks:price to fill in the variable’s information instead of :type :quantitative :title ...

Stream graph

(plot:plot (vega:defplot stream-graph `(:title "Unemployment Stream Graph" :width 300 :height 200 :data (:values ,unemployment-ind) :mark :area :encoding (:x (:field :date :time-unit "yearmonth" :axis (:domain :false :format "%Y" :tick-size 0)) :y (:field count :aggregate :sum :axis null :stack :center) :color (:field :series :scale (:scheme "category20b"))))))

Tabular plots

Table heatmap

(plot:plot (vega:defplot table-heatmap `(:data (:values ,vgcars) :mark :rect :encoding (:x (:field vgcars:cylinders) :y (:field vgcars:origin) :color (:field :horsepower :aggregate :mean)) :config (:axis (:grid t :tick-band :extent)))))

Heatmap with labels

Layering text over a table heatmap

(plot:plot (vega:defplot heatmap-labels `(:data (:values ,vgcars) :transform #((:aggregate #((:op :count :as :num-cars)) :groupby #(:origin :cylinders))) :encoding (:x (:field :cylinders :type :ordinal) :y (:field :origin :type :ordinal)) :layer #((:mark :rect :encoding (:color (:field :num-cars :type :quantitative :title "Count of Records" :legend (:direction :horizontal :gradient-length 120)))) (:mark :text :encoding (:text (:field :num-cars :type :quantitative) :color (:condition (:test "datum['numCars'] < 40" :value :black) :value :white)))) :config (:axis (:grid t :tick-band :extent)))))

Histogram heatmap

(plot:plot (vega:defplot heatmap-histogram `(:data (:values ,imdb) :transform #((:and #((:field :imdb-rating :valid t) (:field :rotten-tomatoes-rating :valid t)))) :mark :rect :width 300 :height 200 :encoding (:x (:bin (:maxbins 60) :field :imdb-rating :type :quantitative :title "IMDB Rating") :y (:bin (:maxbins 40) :field :rotten-tomatoes-rating :type :quantitative :title "Rotten Tomatoes Rating") :color (:aggregate :count :type :quantitative)) :config (:view (:stroke :transparent)))))

Circular plots

Pie chart

(plot:plot (vega:defplot pie-chart `(:data (:values ,(plist-df `(:category ,(aops:linspace 1 6 6) :value #(4 6 10 3 7 8)))) :mark :arc :encoding (:theta (:field :value :type :quantitative) :color (:field :category :type :nominal)))))

Donut chart

(plot:plot (vega:defplot donut-chart `(:data (:values ,(plist-df `(:category ,(aops:linspace 1 6 6) :value #(4 6 10 3 7 8)))) :mark (:type :arc :inner-radius 50) :encoding (:theta (:field :value :type :quantitative) :color (:field :category :type :nominal)))))

Radial plot

This radial plot uses both angular and radial extent to convey multiple dimensions of data. However, this approach is not perceptually effective, as viewers will most likely be drawn to the total area of the shape, conflating the two dimensions. This example also demonstrates a way to add labels to circular plots.

(plot:plot (vega:defplot radial-plot `(:data (:values ,(plist-df '(:value #(12 23 47 6 52 19)))) :layer #((:mark (:type :arc :inner-radius 20 :stroke "#fff")) (:mark (:type :text :radius-offset 10) :encoding (:text (:field :value :type :quantitative)))) :encoding (:theta (:field :value :type :quantitative :stack t) :radius (:field :value :scale (:type :sqrt :zero t :range-min 20)) :color (:field :value :type :nominal :legend nil)))))

Transformations

Normally data transformations should be done in Lisp-Stat with a data frame. These examples illustrate how to accomplish transformations using Vega-Lite. This might be useful if, for example, you’re serving up a lot of plots and want to move the processing to the users browser.

Difference from avg

(plot:plot (vega:defplot difference-from-average `(:data (:values ,(filter-rows imdb '(not (eql imdb-rating :na)))) :transform #((:joinaggregate #((:op :mean ;we could do this above using alexandria:thread-first :field :imdb-rating :as :average-rating))) (:filter "(datum['imdbRating'] - datum.averageRating) > 2.5")) :layer #((:mark :bar :encoding (:x (:field :imdb-rating :type :quantitative :title "IMDB Rating") :y (:field :title :type :ordinal :title "Title"))) (:mark (:type :rule :color "red") :encoding (:x (:aggregate :average :field :average-rating :type :quantitative)))))))

Frequency distribution

Cumulative frequency distribution of films in the IMDB database.

(plot:plot (vega:defplot cumulative-frequency-distribution `(:data (:values ,imdb) :transform #((:sort #((:field :imdb-rating)) :window #((:op :count :field :count as :cumulative-count)) :frame #(nil 0))) :mark :area :encoding (:x (:field :imdb-rating :type :quantitative) :y (:field :cumulative-count :type :quantitative)))))

Layered & cumulative histogram

(plot:plot (vega:defplot layered-histogram `(:data (:values ,(filter-rows imdb '(not (eql imdb-rating :na)))) :transform #((:bin t :field :imdb-rating :as #(:bin-imdb-rating :bin-imdb-rating-end)) (:aggregate #((:op :count :as :count)) :groupby #(:bin-imdb-rating :bin-imdb-rating-end)) (:sort #((:field :bin-imdb-rating)) :window #((:op :sum :field :count :as :cumulative-count)) :frame #(nil 0))) :encoding (:x (:field :bin-imdb-rating :type :quantitative :scale (:zero :false) :title "IMDB Rating") :x2 (:field :bin-imdb-rating-end)) :layer #((:mark :bar :encoding (:y (:field :cumulative-count :type :quantitative :title "Cumulative Count"))) (:mark (:type :bar :color "yellow" :opacity 0.5) :encoding (:y (:field :count :type :quantitative :title "Count")))))))

Layering averages

Layering averages over raw values.

(plot:plot (vega:defplot layered-averages `(:data (:values ,(filter-rows stocks '(string= symbol "GOOG"))) :layer #((:mark (:type :point :opacity 0.3) :encoding (:x (:field :date :time-unit :year) :y (:field :price :type quantitative))) (:mark :line :encoding (:x (:field :date :time-unit :year) :y (:field :price :aggregate :mean)))))))

Error bars

Confidence interval

Error bars showing confidence intervals.

(plot:plot (vega:defplot error-bar-ci `(:data (:values ,barley) :encoding (:y (:field :variety :type :ordinal :title "Variety")) :layer #((:mark (:type :point :filled t) :encoding (:x (:field :yield :aggregate :mean :type :quantitative :scale (:zero :false) :title "Barley Yield") :color (:value "black"))) (:mark (:type :errorbar :extent :ci) :encoding (:x (:field :yield :type :quantitative :title "Barley Yield")))))))

Standard deviation

Error bars showing standard deviation.

(plot:plot (vega:defplot error-bar-sd `(:data (:values ,barley) :encoding (:y (:field :variety :type :ordinal :title "Variety")) :layer #((:mark (:type :point :filled t) :encoding (:x (:field :yield :aggregate :mean :type :quantitative :scale (:zero :false) :title "Barley Yield") :color (:value "black"))) (:mark (:type :errorbar :extent :stdev) :encoding (:x (:field :yield :type :quantitative :title "Barley Yield")))))))

Box plots

Min/max whiskers

A vertical box plot showing median, min, and max body mass of penguins.

(plot:plot (vega:defplot box-plot-min-max `(:data (:values ,penguins) :mark (:type :boxplot :extent "min-max") :encoding (:x (:field :species :type :nominal :title "Species") :y (:field |BODY-MASS-(G)| :type :quantitative :scale (:zero :false) :title "Body Mass (g)") :color (:field :species :type :nominal :legend nil)))))

Tukey

A vertical box plot showing median and lower and upper quartiles of the distribution of body mass of penguins.

(plot:plot (vega:defplot box-plot-tukey `(:data (:values ,penguins) :mark :boxplot :encoding (:x (:field :species :type :nominal :title "Species") :y (:field |BODY-MASS-(G)| :type :quantitative :scale (:zero :false) :title "Body Mass (g)") :color (:field :species :type :nominal :legend nil)))))

Summaries

Box plot with pre-computed summaries. Use this pattern to plot summaries done in a data-frame.

(plot:plot (vega:defplot box-plot-summaries `(:title "Body Mass of Penguin Species (g)" :data (:values ,(plist-df '(:species #("Adelie" "Chinstrap" "Gentoo") :lower #(2850 2700 3950) :q1 #(3350 3487.5 4700) :median #(3700 3700 5000) :q3 #(4000 3950 5500) :upper #(4775 4800 6300) :outliers #(#() #(2700 4800) #())))) :encoding (:y (:field :species :type :nominal :title null)) :layer #((:mark (:type :rule) :encoding (:x (:field :lower :type :quantitative :scale (:zero :false) :title null) :x2 (:field :upper))) (:mark (:type :bar :size 14) :encoding (:x (:field :q1 :type :quantitative) :x2 (:field :q3) :color (:field :species :type :nominal :legend null))) (:mark (:type :tick :color :white :size 14) :encoding (:x (:field :median :type :quantitative))) (:transform #((:flatten #(:outliers))) :mark (:type :point :style "boxplot-outliers") :encoding (:x (:field :outliers :type :quantitative)))))))

Layered

Rolling average

Plot showing a 30 day rolling average with raw values in the background.

(plot:plot (vega:defplot moving-average `(:width 400 :height 300 :data (:values ,seattle-weather) :transform #((:window #((:field :temp-max :op :mean :as :rolling-mean)) :frame #(-15 15))) :encoding (:x (:field :date :type :temporal :title "Date") :y (:type :quantitative :axis (:title "Max Temperature and Rolling Mean"))) :layer #((:mark (:type :point :opacity 0.3) :encoding (:y (:field :temp-max :title "Max Temperature"))) (:mark (:type :line :color "red" :size 3) :encoding (:y (:field :rolling-mean :title "Rolling Mean of Max Temperature")))))))

Histogram w/mean

(plot:plot (vega:defplot histogram-with-mean `(:data (:values ,imdb) :layer #((:mark :bar :encoding (:x (:field :imdb-rating :bin t :title "IMDB Rating") :y (:aggregate :count))) (:mark :rule :encoding (:x (:field :imdb-rating :aggregate :mean :title "Mean of IMDB Rating") :color (:value "red") :size (:value 5)))))))

Interactive

This section demonstrates interactive plots.

Scatter plot matrix

This Vega-Lite interactive scatter plot matrix includes interactive elements and demonstrates creating a SPLOM (scatter plot matrix).

(defparameter vgcars-splom (vega::make-plot "vgcars-splom" vgcars `("$schema" "https://vega.github.io/schema/vega-lite/v5.json" :title "Scatterplot Matrix for Vega Cars" :repeat (:row #(:horsepower :acceleration :miles-per-gallon) :column #(:miles-per-gallon :acceleration :horsepower)) :spec (:data (:url "/data/vgcars-splom-data.json") :mark :point :params #((:name "brush" :select (:type "interval" :resolve "union" :on "[mousedown[event.shiftKey], window:mouseup] > window:mousemove!" :translate "[mousedown[event.shiftKey], window:mouseup] > window:mousemove!" :zoom "wheel![event.shiftKey]")) (:name "grid" :select (:type "interval" :resolve "global" :translate "[mousedown[!event.shiftKey], window:mouseup] > window:mousemove!" :zoom "wheel![!event.shiftKey]") :bind :scales)) :encoding (:x (:field (:repeat "column") :type "quantitative") :y (:field (:repeat "row") :type "quantitative" :axis ("minExtent" 30)) :color (:condition (:param "brush" :field :origin :type "nominal") :value "grey")))))) (plot:plot vgcars-splom)

This example is one of those mentioned in the plotting tutorial that uses a non-standard location for the data property.

Weather exploration

This graph shows an interactive view of Seattle’s weather, including maximum temperature, amount of precipitation, and type of weather. By clicking and dragging on the scatter plot, you can see the proportion of days in that range that have sun, rain, fog, snow, etc.

(plot:plot (vega:defplot weather-exploration `(:title "Seattle Weather, 2012-2015" :data (:values ,seattle-weather) :vconcat #(;; upper graph (:encoding (:color (:condition (:param :brush :title "Weather" :field :weather :type :nominal :scale (:domain #("sun" "fog" "drizzle" "rain" "snow") :range #("#e7ba52", "#a7a7a7", "#aec7e8", "#1f77b4", "#9467bd"))) :value "lightgray") :size (:field :precipitation :type :quantitative :title "Precipitation" :scale (:domain #(-1 50))) :x (:field :date :time-unit :monthdate :title "Date" :axis (:format "%b")) :y (:field :temp-max :type :quantitative :scale (:domain #(-5 40)) :title "Maximum Daily Temperature (C)")) :width 600 :height 300 :mark :point :params #((:name :brush :select (:type :interval :encodings #(:x)))) :transform #((:filter (:param :click)))) ;; lower graph (:encoding (:color (:condition (:param :click :field :weather :scale (:domain #("sun", "fog", "drizzle", "rain", "snow") :range #("#e7ba52", "#a7a7a7", "#aec7e8", "#1f77b4", "#9467bd"))) :value "lightgray") :x (:aggregate :count) :y (:field :weather :title "Weather")) :width 600 :mark :bar :params #((:name :click :select (:type :point :encodings #(:color)))) :transform #((:filter (:param :brush))))))))

Interactive scatterplot

(plot:plot (vega:defplot global-health `(:title "Global Health Statistics by Country and Year" :data (:values ,gapminder) :width 800 :height 500 :layer #((:transform #((:filter (:field :country :equal "afghanistan")) (:filter (:param :year))) :mark (:type :text :font-size 100 :x 420 :y 250 :opacity 0.06) :encoding (:text (:field :year))) (:transform #((:lookup :cluster :from (:key :id :fields #(:name) :data (:values #(("id" 0 "name" "South Asia") ("id" 1 "name" "Europe & Central Asia") ("id" 2 "name" "Sub-Saharan Africa") ("id" 3 "name" "America") ("id" 4 "name" "East Asia & Pacific") ("id" 5 "name" "Middle East & North Africa")))))) :encoding (:x (:field :fertility :type :quantitative :scale (:domain #(0 9)) :axis (:tick-count 5 :title "Fertility")) :y (:field :life-expect :type :quantitative :scale (:domain #(20 85)) :axis (:tick-count 5 :title "Life Expectancy"))) :layer #((:mark (:type :line :size 4 :color "lightgray" :stroke-cap "round") :encoding (:detail (:field :country) :order (:field :year) :opacity (:condition (:test (:or #((:param :hovered :empty :false) (:param :clicked :empty :false))) :value 0.8) :value 0))) (:params #((:name :year :value #((:year 1955)) :select (:type :point :fields #(:year)) :bind (:name :year :input :range :min 1955 :max 2005 :step 5)) (:name :hovered :select (:type :point :fields #(:country) :toggle :false :on :mouseover)) (:name :clicked :select (:type :point :fields #(:country)))) :transform #((:filter (:param :year))) :mark (:type :circle :size 100 :opacity 0.9) :encoding (:color (:field :name :title "Region"))) (:transform #((:filter (:and #((:param :year) (:or #((:param :clicked :empty :false) (:param :hovered :empty :false))))))) :mark (:type :text :y-offset -12 :font-size 12 :font-weight :bold) :encoding (:text (:field :country) :color (:field :name :title "Region"))) (:transform #((:filter (:param :hovered :empty :false)) (:filter (:not (:param :year)))) :layer #((:mark (:type :text :y-offset -12 :font-size 12 :color "gray") :encoding (:text (:field :year))) (:mark (:type :circle :color "gray"))))))))))

Crossfilter

Cross-filtering makes it easier and more intuitive for viewers of a plot to interact with the data and understand how one metric affects another. With cross-filtering, you can click a data point in one dashboard view to have all dashboard views automatically filter on that value.

Click and drag across one of the charts to see the other variables filtered.

(plot:plot (vega:defplot cross-filter `(:title "Cross filtering of flights" :data (:values ,flights-2k) :transform #((:calculate "hours(datum.date)", :as "time")) ;what does 'hours' do? :repeat (:column #(:distance :delay :time)) :spec (:layer #((:params #((:name :brush :select (:type :interval :encodings #(:x)))) :mark :bar :encoding (:x (:field (:repeat :column) :bin (:maxbins 20)) :y (:aggregate :count) :color (:value "#ddd"))) (:transform #((:filter (:param :brush))) :mark :bar :encoding (:x (:field (:repeat :column) :bin (:maxbins 20)) :y (:aggregate :count))))))))