{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Draws a Correlation Chart or Heatmap on the Titanic data set"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### NuGet package installation"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"Installing package MathNet.Numerics, version 4.9.0.....................................done!"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Successfully added reference to package MathNet.Numerics, version 4.9.0"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Installing package Microsoft.ML, version 1.4.0.........done!"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Successfully added reference to package Microsoft.ML, version 1.4.0"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Installing package XPlot.Plotly, version 3.0.1........done!"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Successfully added reference to package XPlot.Plotly, version 3.0.1"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"#r \"nuget:MathNet.Numerics, 4.9.0\"\n",
"#r \"nuget:Microsoft.ML, 1.4.0\"\n",
"#r \"nuget:XPlot.Plotly, 3.0.1\""
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Namespaces"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"using Microsoft.ML;\n",
"using Microsoft.ML.Data;\n",
"using XPlot.Plotly;\n",
"using MathNet.Numerics.Statistics;"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Simple heatmap sample\n",
"Warming up ..."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"var graph = new Graph.Heatmap()\n",
"{\n",
" x = new [] { \"one\", \"two\", \"three\"},\n",
" y = new [] { \"three\", \"two\", \"one\" },\n",
" z = new List> \n",
" { \n",
" new List { 0, -.75, 1 }, \n",
" new List { .75, 1, -.75 }, \n",
" new List { 1, .75, 0 }\n",
" },\n",
" zmin = -1,\n",
" zmax = 1\n",
"};\n",
"\n",
"var chart = Chart.Plot(graph);\n",
"\n",
"var layout = new Layout.Layout(){ title=\"Sample Correlation Chart\" };\n",
"chart.WithLayout(layout);\n",
"\n",
"display(chart);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## And now for the real thing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Read the data"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"var mlContext = new MLContext(seed: null);\n",
"\n",
"var readerOptions = new TextLoader.Options()\n",
"{\n",
" Separators = new[] { ',' },\n",
" HasHeader = true,\n",
" AllowQuoting = true,\n",
" Columns = new[]\n",
" {\n",
" new TextLoader.Column(\"Survived\", DataKind.Single, 1),\n",
" new TextLoader.Column(\"PClass\", DataKind.Single, 2),\n",
" new TextLoader.Column(\"Age\", DataKind.Single, 5),\n",
" new TextLoader.Column(\"SibSp\", DataKind.Single, 6),\n",
" new TextLoader.Column(\"Parch\", DataKind.Single, 7),\n",
" new TextLoader.Column(\"Fare\", DataKind.Single, 9)\n",
" }\n",
"};\n",
"\n",
"var dataView = mlContext.Data.LoadFromTextFile(\"./Titanic.csv\", readerOptions);"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Enter the Matrix\n",
"\n",
"ML.NET prefers singles (float), Math.NET prefers doubles ..."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"var matrix = new List>();\n",
"for (int i = 0; i < dataView.Schema.Count; i++)\n",
"{\n",
" var column = dataView.Schema[i];\n",
" matrix.Add(dataView.GetColumn(column).Select(f => (double)f).ToList());\n",
"}\n",
"\n",
"var data = new double[6, 6];\n",
"for (int x = 0; x < 6; ++x)\n",
"{\n",
" for (int y = 0; y < 5 - x; ++y)\n",
" {\n",
" var seriesA = matrix[x];\n",
" var seriesB = matrix[5 - y];\n",
"\n",
" var value = Correlation.Pearson(seriesA, seriesB);\n",
"\n",
" data[x, y] = value;\n",
" data[5 - y, 5 - x] = value;\n",
" }\n",
"\n",
" data[x, 5 - x] = 1;\n",
"}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Draw"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"var graph = new Graph.Heatmap()\n",
"{\n",
" y = new [] { \"Survived\", \"Class\", \"Age\", \"Sib / Sp\", \"Par / Chi\", \"Fare\" },\n",
" x = new [] { \"Fare\", \"Parents / Children\", \"Siblings / Spouses\", \"Age\", \"Class\", \"Survived\" },\n",
" z = data,\n",
" zmin = -1,\n",
" zmax = 1\n",
"};\n",
"\n",
"var chart = Chart.Plot(graph);\n",
"\n",
"var layout = new Layout.Layout(){ title=\"Titanic Survival Correlation Chart\" };\n",
"chart.WithLayout(layout);\n",
"\n",
"display(chart);"
]
}
],
"metadata": {
"kernelspec": {
"display_name": ".NET (C#)",
"language": "C#",
"name": ".net-csharp"
},
"language_info": {
"file_extension": ".cs",
"mimetype": "text/x-csharp",
"name": "C#",
"pygments_lexer": "csharp",
"version": "8.0"
}
},
"nbformat": 4,
"nbformat_minor": 2
}