Super basic qplot implementation on Scala collection, using R backend as renderer... in 74 lines :-)

QPlot.snb

{
  "metadata" : {
    "name" : "QPlot",
    "user_save_timestamp" : "1970-01-01T01:00:00.000Z",
    "auto_save_timestamp" : "1970-01-01T01:00:00.000Z",
    "language_info" : {
      "name" : "scala",
      "file_extension" : "scala",
      "codemirror_mode" : "text/x-scala"
    },
    "trusted" : true,
    "customLocalRepo" : null,
    "customRepos" : null,
    "customDeps" : null,
    "customImports" : null,
    "customArgs" : null,
    "customSparkConf" : null
  },
  "cells" : [ {
    "metadata" : {
      "id" : "2D571E3F82424F4080D5BB2E222D57C1"
    },
    "cell_type" : "markdown",
    "source" : "# QPlot"
  }, {
    "metadata" : {
      "id" : "8DFC9A1BAD974742BE5B1E25828F19F7"
    },
    "cell_type" : "markdown",
    "source" : "> Warning:\n>\n> You'll need `R` install locally, \n> you'll also need `ggplot2` \n> * ```install.packages(\"ggplot2\")```\n> * or on ubuntu, ```apt-get install r-cran-ggplot2```.\n>"
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "106C73B4ED4A47E4AC31C2176A5467B9"
    },
    "cell_type" : "code",
    "source" : "import scala.language.postfixOps\nimport SamplerImplicits._\ndef qplot[C:ToPoints:Sampler](\n  c:C, \n  x:String=\"\", y:String=\"\", \n  geom:List[String]=Nil, fill:String=\"\", \n  method:String=\"\", formula:String=\"\", color:String=\"\",\n  main:String=\"\", xlab:String=\"\", ylab:String=\"\",\n  width:String=\"150px\", height:String=\"150px\"\n) = {\n  val toPoints = implicitly[ToPoints[C]]\n  val points = toPoints(c, toPoints.count(c).toInt)\n  val ps = points.head.headers.zipWithIndex.map { case (h, i) =>\n    h → points.map(_.values(i))\n  }\n  val cols = ps.map { case (h, vs) =>\n  val arr = vs.map(v => toJson(v).toString)\n    s\"\"\"\n     $h <- c(${arr.mkString(\",\")})\n    \"\"\"\n  }.mkString(\"\\n\")\n  val r = s\"\"\"\n  $cols\n\n  df <- data.frame(${ps.map(_._1).mkString(\",\")})\n  \"\"\"\n  import sys.process._\n  \n  val scriptFile = java.io.File.createTempFile(\"r-script\", \".R\")\n  val img = java.io.File.createTempFile(\"r-plot\", \".png\")\n  \n  def addString(param:String, s:String) = s match {\n    case \"\" => \"\"\n    case x  => s\"\"\", $param = \"$s\" \"\"\"\n  }\n  \n  def addField(param:String, s:String) = s match {\n    case \"\" => \"\"\n    case x  => s\"\"\", $param = $s \"\"\"\n  }\n  \n  val plot = s\"\"\"\n  library(ggplot2) \n\n\n  q <- qplot(data=df\n             ${addField(\"x\", x)}\n             ${addField(\"y\", y)}\n             ,geom=c(${geom.map(x => \"\\\"\"+x+\"\\\"\").mkString(\",\")}) \n             ${addField(\"fill\", fill)}, \n             alpha=I(.5), \n             main=\"$main\", xlab=\"$xlab\", ylab=\"$ylab\"\n             ${addString(\"method\", method)}\n             ${addField(\"formula\", formula)}\n             ${addField(\"color\", color)}\n             )\n  ggsave(\"${img.getAbsolutePath}\", plot=q)\n  dev.off()\n  \"\"\"\n  val script = s\"\"\"\n  $r\n\n  $plot\n  \"\"\"\n  val w = new java.io.FileWriter(scriptFile)\n  w.write(script)\n  w.close()\n  \n  Process(s\"Rscript ${scriptFile.getAbsolutePath}\") !!\n  \n  val i = widgets.img(\"png\", width, height)\n  i.file(img)\n  i\n}",
    "outputs" : [ ]
  }, {
    "metadata" : {
      "id" : "0484E38D76AF4EFA87B9582587728FE7"
    },
    "cell_type" : "markdown",
    "source" : "#Cars"
  }, {
    "metadata" : {
      "id" : "281343A362F3495399844B78225EC361"
    },
    "cell_type" : "markdown",
    "source" : "## Data"
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "E39FD8741B1C4992985C084AE6CB72BE"
    },
    "cell_type" : "code",
    "source" : "case class Car(model:String, mpg:Double, cyl:String, disp:Double, hp:Double, drat:Double, wt:Double, qsec:Double, vs:Double, am:String, gear:String, carb:Double)",
    "outputs" : [ ]
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "23484468D5E745A88AC771249EA3FF3B"
    },
    "cell_type" : "code",
    "source" : ":sh wget https://gist.githubusercontent.com/seankross/a412dfbd88b3db70b74b/raw/5f23f993cd87c283ce766e7ac6b329ee7cc2e1d1/mtcars.csv -O /tmp/mtcars.csv",
    "outputs" : [ ]
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "45AA664E4979497B82DDA14B0E4A767D"
    },
    "cell_type" : "code",
    "source" : "val cars = {\n  val headers :: lines = scala.io.Source.fromFile(\"/tmp/mtcars.csv\").getLines.toList\n  \n  lines.map(_.split(\",\").toList).map{case model::mpg::cyl::disp::hp::drat::wt::qsec::vs::am::gear::carb::Nil => \n                                             Car(model,mpg.toDouble,cyl+\"cyl\",disp.toDouble,hp.toDouble,drat.toDouble,wt.toDouble,\n                                                 qsec.toDouble,vs.toDouble,if(am.toInt==1)\"Manual\"else\"Automatic\",gear+\"gears\",carb.toDouble)\n                                              case _ => ???\n                                            }\n}",
    "outputs" : [ ]
  }, {
    "metadata" : {
      "id" : "6CE76626C59340C59C9CC6ECE201598A"
    },
    "cell_type" : "markdown",
    "source" : "## Plots"
  }, {
    "metadata" : {
      "id" : "A9C2E8E7C91343868AC1E50582DEFF49"
    },
    "cell_type" : "markdown",
    "source" : "Based on the example page on [qplot](http://www.statmethods.net/advgraphs/ggplot2.html)."
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "D21749B5D65C4AE8819E68FF8457FB22"
    },
    "cell_type" : "code",
    "source" : "qplot(cars, x=\"mpg\", geom=List(\"density\"), fill=\"gear\", main=\"Distribution of Gas Milage\", xlab=\"Miles Per Gallon\", ylab=\"Density\")",
    "outputs" : [ ]
  }, {
    "metadata" : {
      "trusted" : true,
      "input_collapsed" : false,
      "collapsed" : false,
      "id" : "0A2A8CCF1B614B068E369219049886D6"
    },
    "cell_type" : "code",
    "source" : "qplot(cars, x=\"wt\", y=\"mpg\", geom=List(\"point\", \"smooth\"), \n   method=\"lm\", formula=\"y~x\", color=\"cyl\", \n   main=\"Regression of MPG on Weight\", \n   xlab=\"Weight\", ylab=\"Miles per Gallon\",\n   width=\"500px\", height=\"500px\")",
    "outputs" : [ ]
  } ],
  "nbformat" : 4
}