MoneyballExample.java

package com.github.lwhite1.tablesaw.api.ml.regression;

import com.github.lwhite1.tablesaw.api.IntColumn;
import com.github.lwhite1.tablesaw.api.NumericColumn;
import com.github.lwhite1.tablesaw.api.Table;
import com.github.lwhite1.tablesaw.api.plot.Histogram;
import com.github.lwhite1.tablesaw.api.plot.Scatter;
import com.github.lwhite1.tablesaw.columns.Column;

import static com.github.lwhite1.tablesaw.api.QueryHelper.column;

/**
 * An example doing ordinary least squares regression
 */
public class MoneyballExample {

    public static void main(String[] args) throws Exception {

        // Get the data
        Table baseball = Table.createFromCsv("data/baseball.csv");
        out(baseball.structure().print());

        // filter to the data available in the 2002 season
        Table moneyball = baseball.selectWhere(column("year").isLessThan(2002));

        // plot regular season wins against year, segregating on whether the team made the plays
        NumericColumn wins = moneyball.numericColumn("W");
        NumericColumn year = moneyball.numericColumn("Year");
        Column playoffs = moneyball.column("Playoffs");
        Scatter.show("Regular season wins by year", wins, year, moneyball.splitOn(playoffs));

        // Calculate the run difference for use in the regression model
        IntColumn runDifference = moneyball.shortColumn("RS").subtract(moneyball.shortColumn("RA"));
        moneyball.addColumn(runDifference);
        runDifference.setName("RD");

        // Plot RD vs Wins to see if the relationship looks linear
        Scatter.show("RD x Wins", moneyball.numericColumn("RD"), moneyball.numericColumn("W"));

        // Create the regression model
        //ShortColumn wins = moneyball.shortColumn("W");
        LeastSquares winsModel = LeastSquares.train(wins, runDifference);
        out(winsModel);

        // Make a prediction of how many games we win if we score 135 more runs than our opponents
        double[] testValue = new double[1];
        testValue[0] = 135;
        double prediction = winsModel.predict(testValue);
        out("Predicted wins with RD = 135: " + prediction);

        // Predict runsScored based on On-base percentage, batting average and slugging percentage

        LeastSquares runsScored = LeastSquares.train(moneyball.nCol("RS"),
                moneyball.nCol("OBP"), moneyball.nCol("BA"), moneyball.nCol("SLG"));
        out(runsScored);

        LeastSquares runsScored2 = LeastSquares.train(moneyball.nCol("RS"),
                moneyball.nCol("OBP"), moneyball.nCol("SLG"));
        out(runsScored2);

        Histogram.show(runsScored2.residuals());

        Scatter.fittedVsResidual(runsScored2);
        Scatter.actualVsFitted(runsScored2);

        // We use opponent OBP and opponent SLG to model the efficacy of our pitching and defence

        Table moneyball2 = moneyball.selectWhere(column("year").isGreaterThan(1998));
        LeastSquares runsAllowed = LeastSquares.train(moneyball2.nCol("RA"),
                moneyball2.nCol("OOBP"), moneyball2.nCol("OSLG"));
        out(runsAllowed);

    }

    private static void out(Object o) {
        System.out.println(String.valueOf(o));
    }
}