Skip to content

PlotGuide.md

Latest commit

 

History

History
executable file
·
199 lines (155 loc) · 6.58 KB

PlotGuide.md

File metadata and controls

executable file
·
199 lines (155 loc) · 6.58 KB

uni.plot — Plot Guide

uni.plot adds interactive and file-based charting directly on MatD via XChart (GGPlot2 theme).

import uni.data.*
import uni.plot.*

Methods

.pairs() — scatterplot matrix

A p×p grid of subplots: histograms on the diagonal, scatter plots off-diagonal. Each scatter cell shows axis tick marks, gridlines, and labelled x/y variable names. Equivalent to R pairs() / seaborn pairplot.

The window is resizable — the plot re-renders at the new size automatically.

m.pairs(
  title        = "Iris: scatterplot matrix",  // window title
  labels       = Seq("sepal_l", "sepal_w", "petal_l", "petal_w"),
  bins         = 10,                       // histogram bucket count (default 10)
  dotSize      = 9,                        // scatter dot size in pixels (default 3)
  scatterAlpha = 90,                       // dot opacity 0–255 (default 80; lower = more transparent)
  color        = Color(31, 119, 180),      // bar and dot colour (default blue)
  labelStyle   = Font.BOLD,                // Font.PLAIN / Font.BOLD / Font.ITALIC (default BOLD)
  saveTo       = "docs/images/iris-pairs", // saves iris-pairs.png; omit to show window
  style        = PlotStyle(width = 1500, height = 900),
)

scatterAlpha is worth tuning: low values (40–60) reveal density in crowded plots; higher values (150+) make sparse plots easier to read.

labelStyle accepts any java.awt.Font style constant — import java.awt.Font to use them.


Fisher Iris — 4-feature pairplot (150 samples): note the strong petal_l ↔ petal_w correlation and the near-zero sepal_w correlation

.plot() — line chart

One series per column; x-axis = row indices.

m.plot(
  title  = "my chart",               // window / PNG title
  labels = Seq("col0", "col1"),      // series labels (defaults: "col0", "col1", …)
  saveTo = "out/chart",              // saves out/chart.png; omit to show window
  style  = PlotStyle(width = 900, height = 600, xLabel = "time", yLabel = "value"),
)

.scatter() — scatter plot

Two columns plotted against each other; optional group colouring.

m.scatter(
  xCol     = 0,                      // column index for x-axis (default 0)
  yCol     = 1,                      // column index for y-axis (default 1)
  groupCol = 4,                      // colour points by this column (-1 = no grouping)
  title    = "x vs y",
  saveTo   = "out/scatter",
  style    = PlotStyle(width = 700, height = 700, xLabel = "feature A", yLabel = "feature B"),
)

.hist() — histogram

All values in the matrix binned and plotted as an area chart.

m.hist(
  bins   = 20,                       // number of bins (default 20)
  title  = "distribution",
  saveTo = "out/hist",
  style  = PlotStyle(width = 800, height = 500, yLabel = "frequency"),
)

.bar() — bar chart

One bar per row; labels come from a column or default to row indices.

m.bar(
  col      = 1,                      // column providing bar heights (default 0)
  labelCol = 0,                      // column providing x-axis labels (-1 = row indices)
  title    = "totals",
  saveTo   = "out/bar",
  style    = PlotStyle(xLabel = "category", yLabel = "count"),
)

.heatmap() — heatmap

Renders the matrix as a colour grid. Useful for correlation and confusion matrices.

val features = Seq("sepal_l", "sepal_w", "petal_l", "petal_w")
corr.heatmap(
  title     = "Correlation matrix",
  rowLabels = features,
  colLabels = features,
  saveTo    = "out/corr",
  style     = PlotStyle(width = 700, height = 700),
)

.boxPlot() — box plot

One box per column showing median, quartiles, and outliers.

m.boxPlot(
  title  = "feature distributions",
  labels = Seq("sepal_l", "sepal_w", "petal_l", "petal_w"),
  saveTo = "out/boxes",
  style  = PlotStyle(yLabel = "cm"),
)

PlotStyle

All fields are optional — omit any field to keep the GGPlot2 theme default.

Field Type Default Effect
width Int method-specific chart width in pixels
height Int method-specific chart height in pixels
background Option[Color] None outer chart background
plotBackground Option[Color] None inner plot-area background
foreground Option[Color] None axis labels and title colour
seriesColors Seq[Color] Nil per-series colours in order
xLabel String "" x-axis label
yLabel String "" y-axis label
xLog Boolean false logarithmic x-axis (XY charts only)
yLog Boolean false logarithmic y-axis (XY charts only)

Named presets

Preset Value Use case
PlotStyle.uniform PlotStyle(width = 800, height = 500) Consistent dimensions for saved images

Examples

Axis labels:

m.scatter(style = PlotStyle(xLabel = "principal component 1", yLabel = "principal component 2"))

Log scale:

m.plot(style = PlotStyle(yLog = true, yLabel = "loss (log scale)"))

Dark background:

import java.awt.Color
m.hist(style = PlotStyle(
  background     = Some(Color.BLACK),
  plotBackground = Some(new Color(30, 30, 30)),
  foreground     = Some(Color.WHITE),
))

Custom series colours:

import java.awt.Color
m.plot(style = PlotStyle(
  seriesColors = Seq(Color.RED, Color.BLUE, Color.GREEN),
))

Uniform export size for documentation:

m.scatter(2, 3, saveTo = "docs/images/iris-scatter", style = PlotStyle.uniform)
m.hist(bins = 20, saveTo = "docs/images/iris-hist",  style = PlotStyle.uniform)
// → both PNGs are exactly 800×500

Demo scripts

Script Description
jsrc/corr.sc Interactive: Iris correlation heatmap + two scatter windows showing strong vs weak correlation
jsrc/iris.sc Interactive: scatter, histogram, and line plot on the Fisher Iris dataset
jsrc/anscombe.sc Interactive: Anscombe's Quartet — four scatter plots with identical statistics
jsrc/irisPairedFeatures.sc Interactive: pairs (scatterplot matrix) on the Fisher Iris dataset
jsrc/airfoilNoise.sc Interactive: pairs (scatterplot matrix) on the UCI Airfoil Self-Noise dataset
jsrc/gen-images.sc Headless: regenerates all docs/images/iris-*.png (line, scatter, grouped-scatter, hist, bar, corr, box, pairs)