static async Task Main(string[] args)
{
try
{
var datasetColumns = new DataSetColumns();
PopulateDataSet(datasetColumns);
// Learn about your data
UnderstandData(datasetColumns);
// Find Correlations
//FindCorrelations(datasetColumns);
// Build and train your model
//PredictionModel<TaxiTrip, TaxiTripFarePrediction> model = await Train();
// Evaluate your model (sample test to see how accurate your model is)
//Evaluate(model);
// Use your model.
//TaxiTripFarePrediction prediction = model.Predict(TestTrips.Trip1);
//Console.WriteLine("Predicted fare: {0}, actual fare: 29.5", prediction.FareAmount);
}
catch (Exception e)
{
Console.WriteLine(e);
}
Console.WriteLine("Execution Complete -- Press Enter");
Console.ReadLine();
}
public static void PopulateDataSet(DataSetColumns dataSetColumns)
{
var riders = new List <int>();
var tripTime = new List <int>();
var tripDistance = new List <double>();
var fareAmount = new List <double>();
using (var reader = new StreamReader(_datapath))
{
reader.ReadLine(); // headers
while (!reader.EndOfStream)
{
var line = reader.ReadLine();
var values = line.Split(',');
riders.Add(int.Parse(values[2]));
tripTime.Add(int.Parse(values[3]));
tripDistance.Add(double.Parse(values[4]));
fareAmount.Add(double.Parse(values[6]));
}
}
dataSetColumns.Riders = riders.ToArray();
dataSetColumns.TripTime = tripTime.ToArray();
dataSetColumns.TripDistance = tripDistance.ToArray();
dataSetColumns.FareAmount = fareAmount.ToArray();
}
public static void FindCorrelations(DataSetColumns dataSetColumns)
{
var pcc = PearsonCoefficientCalculator.ComputeCoefficient(dataSetColumns.TripDistance, dataSetColumns.TripTime.Select(i => (double)i).ToArray());
Console.WriteLine("Pearson coefficient between Time and Distance");
Console.WriteLine(pcc);
pcc = PearsonCoefficientCalculator.ComputeCoefficient(dataSetColumns.TripDistance, dataSetColumns.Riders.Select(i => (double)i).ToArray());
Console.WriteLine("Pearson coefficient between Riders and Distance");
Console.WriteLine(pcc);
}
/// <summary>
/// This method reads the training data and calculates a variety of statistical data about each column.
/// </summary>
public static void UnderstandData(DataSetColumns dataSetColumns)
{
//Calculate and output the statistics for your pleasure
var stats = MetricStatistics.CalculateMetricStatisticsAsync(dataSetColumns.Riders, DateTime.Now, TimeSpan.FromSeconds(1)).Result;
Console.WriteLine("Rider Statistics:");
Console.WriteLine(stats.ToString());
stats = MetricStatistics.CalculateMetricStatisticsAsync(dataSetColumns.TripTime, DateTime.Now, TimeSpan.FromSeconds(1)).Result;
Console.WriteLine("Trip Time Statistics:");
Console.WriteLine(stats.ToString());
stats = MetricStatistics.CalculateMetricStatisticsAsync(dataSetColumns.TripDistance, DateTime.Now, TimeSpan.FromSeconds(1)).Result;
Console.WriteLine("Trip Distance Statistics:");
Console.WriteLine(stats.ToString());
stats = MetricStatistics.CalculateMetricStatisticsAsync(dataSetColumns.FareAmount, DateTime.Now, TimeSpan.FromSeconds(1)).Result;
Console.WriteLine("Fare Amount Statistics:");
Console.WriteLine(stats.ToString());
}