How to create a multi-series (line) chart with scala-chart

Question

I came across Scala-chart, a really nice wrapper for using jFreeChart in Scala. It has several utility classes for generating charts with minimal effort.

What is the best way to generate a line chart with multiple series using scala-chart?

Answer 1

In scala-chart there are several different ways to create a multi-series line chart. Which way to use depends on how you create your dataset (including ways to work with legacy JFreeChart code):

• (for comparison) create a single-series line chart:

  val series = for (x <- 1 to 5) yield (x,x*x)
  val chart = XYLineChart(series)
  

• build up a multi-series line chart entirely from scala collections (this way I recommend because it is the most idiomatic):

  val names: List[String] = "Series A" :: "Series B" :: Nil
  
  val data = for {
    name <- names
    series = for (x <- 1 to 5) yield (x,util.Random.nextInt(5))
  } yield name -> series
  
  val chart = XYLineChart(data)
  

• from a collection of XYSeries objects:

  val names: List[String] = "Series A" :: "Series B" :: Nil
  
  def randomSeries(name: String): XYSeries =
    List.tabulate(5)(x => (x,util.Random.nextInt(5))).toXYSeries(name)
  
  val data = for (name <- names) yield randomSeries(name)
  
  val chart = XYLineChart(data)
  

• explicitly create an XYSeriesCollection object:

  def data: XYSeriesCollection = ???
  val chart = XYLineChart(data)
  

These are some simple snippets but they should illustrate how data creation is possible and most of the time it boils down to one of these ways.

---

The current implementation (as of 0.4.0) work in the way that:

• all chart factories accept an arbitrary data object of type A and
• in addition accept an implicit -- a type class like ToXYDataset[A] -- which converts this arbitrary data of type A to the respective dataset the factory is based on
• in the case of xy charts there are default converters for XYSeries, XYSeriesCollection, Coll[XYSeries] (where coll is a scala standard library collection), Coll[(A,B)] and Coll[(A,Coll[(B,C)])]

This way the chart factories are fully extendable and can be used with data instances of your custom types -- you just have to write your own conversion type class instances for your custom types.