public static void Sample(object caller, float scale, string year, int gender,
IList <double> source, IList <double> weights, string title, string xlabel = "age", string ylabel = "population", [CallerMemberName] string callerName = "")
{
var env = new SimSharp.ThreadSafeSimulation(42);
var length = source.Count;
var total = 0;
var simulatedTotal = 0;
var cbsPoints = new List <DataPoint>(length);
for (int i = 0; i < length; i++)
{
total += (int)weights[i];
simulatedTotal += (int)(weights[i] * scale);
cbsPoints.Add(new DataPoint(i, (int)(weights[i] * scale)));
}
int[] data = new int[length];
// Sample some probability data, based on scale of total.
for (var s = 0; s < total * scale; s++)
{
data[(int)env.RandChoice(source,
weights)]++;
}
var points = new List <DataPoint>(length);
for (int j = 0; j < data.Length; j++)
{
points.Add(new DataPoint(j, data[j]));
}
// Assert (through visual inspection)
var model = new PlotModel {
Title = $"{title} {year} - {total} citizens {Constants.DisplayNames[gender]} scale {100 * scale}% "
};
model.DefaultColors = new List <OxyColor> {
OxyColors.Silver, OxyColors.Black
};
model.Axes.Add(new LinearAxis {
Position = AxisPosition.Bottom, Title = xlabel
});
model.Axes.Add(new LinearAxis {
Position = AxisPosition.Left, Title = ylabel
});
model.Series.Add(new LineSeries {
ItemsSource = points, StrokeThickness = 4
});
model.Series.Add(new LineSeries {
ItemsSource = cbsPoints, StrokeThickness = 1
});
model.Series[0].Title = "simulated";
model.Series[1].Title = "actual";
Helpers.SvgWriter(model, caller, $"{year}-{total}-{Constants.DisplayNames[gender]}-scale-{scale}", new[] { points, cbsPoints }, callerName);
}
// TODO: This probability data set does not contain years. Year not specified.
public void MaritalAge(float scale, int startDate, int endDate, int gender)
{
var env = new SimSharp.ThreadSafeSimulation(42);
for (int y = startDate; y <= endDate; y++)
{
var index = MaritalStatusProbability.YearIndex(y);
Helpers.Sample(this, scale, y.ToString(), gender,
MaritalStatusProbability.MaritalAgeSource,
MaritalStatusProbability.MaritalAgeWeights[gender, index],
"Marital Age distribution"
);
}
}
public void AgeDistribution(float scale, int startYear, int endYear, int gender)
{
// Arrange
var env = new SimSharp.ThreadSafeSimulation(42);
for (int y = startYear; y <= endYear; y++)
{
var index = AgeProbability.YearIndex(y);
Helpers.Sample(this, scale, y.ToString(), gender,
AgeProbability.Source[gender, index],
AgeProbability.Weights[gender, index],
"Age distribution"
);
}
}
public static void MaritalDuration1()
{
//Arrange
var env = new SimSharp.ThreadSafeSimulation(42);
var collection = new MaritalDurationCollection();
var coupleList = new List <double>();
//Act
//The married couples are randomly assigned into a marital duration list, using the duration weights given in the "MaritalDuration" class
for (int i = 0; i < 3217582; i++)
{
coupleList.Add(env.RandChoice(DivorceProbability.DivorceRateSource, DivorceProbability.DivorceRateWeights));
}
//The couples are sorted into a collection based on their i variable.
// i is in this case the years someone is married, the duration.
for (double i = 0; i < 72; i++)
{
collection.Add(new MaritalDurationProbability()
{
Duration = i,
MarriedDurationCouples = coupleList.Where(d => d == i).Count()
});
}
//Collection is exported into a CSV file saved in the location below.
var export = Frame.FromRecords(collection);
export.SaveCsv($"{Global.GetDataFolder()}maritalDurationList.csv");
//Assert
//We took 3 random durations from the list and we asserted them to see if they are
//really the same number each time.
//We tested it to 5 years, 31 years and, 54 years.
int actualDuration5 = coupleList.Where(d => d == 5).Count();
int expectedDuration5 = 59944;
int actualDuration54 = coupleList.Where(d => d == 54).Count();
int expectedDuration54 = 34631;
int actualDuration31 = coupleList.Where(d => d == 31).Count();
int expectedDuration31 = 53246;
Assert.Equal(expectedDuration5, actualDuration5);
Assert.Equal(expectedDuration31, actualDuration31);
Assert.Equal(expectedDuration54, actualDuration54);
}
public static void ReadFile()
{
// Arrange
Frame <int, string> frame;
List <double> weights;
var collection = new MaritalStatusProbabilityCollection();
frame = Frame.ReadCsv("../../../../../doc/data/marital_status.csv");
var env = new SimSharp.ThreadSafeSimulation(42);
// Act
//For the years 1950 to 2019 we are going to put a test amount of people through the weights and see what happens
for (int j = 1950; j < 2020; j++)
{
//A list partnertype made, where we will put everyone in.
var people = new List <PartnerType>();
//Weights are decided from "marital_status.csv", each year takes their own amount of weights.
weights = frame.GetColumn <double>(Convert.ToString(j)).Values.Select(c => Convert.ToDouble(c)).Take(3).ToList();
MaritalStatus.Weights = weights;
//For 100000 citizens of each year, they each get randomly sorted in a source, based on the
//weights given above.
for (int i = 0; i < 100000; i++)
{
people.Add(env.RandChoice(MaritalStatus.Source, MaritalStatus.Weights));
}
//Here we are added all of them into a collection based on the MaritalStatusPorbability class which can be viewed at the top.
collection.Add(new MaritalStatusProbability()
{
Singles = people.Where(p => p == PartnerType.Single).Count(),
Married = people.Where(p => p == PartnerType.Married).Count(),
Partnership = people.Where(p => p == PartnerType.Partnership).Count(),
Year = j
});
}
//This gets exported into a CSV file which can be located as stated below.
var export = Frame.FromRecords(collection);
export.SaveCsv($"{Global.GetDataFolder()}maritalStatusWeights.csv");
}
public void ChildrenProbabilities()
{
//Arrange
var env = new SimSharp.ThreadSafeSimulation(42);
Frame <int, string> frame;
frame = Frame.ReadCsv($"{Global.GetDataFolder()}maritalAgeProbability.csv");
var womenAges = frame.GetColumn <double>("MarriedWomen").Values.Select(c => Convert.ToDouble(c)).ToList();
Frame <int, string> MotherData = Frame.ReadCsv("../../../../../doc/data/motherbirthing_ages.csv");
List <double> weights;
List <double> childChoice;
var childAmount = new Dictionary <int, List <double> >();
var collection = new ChildrenProbabilityCollection();
var collectionNew = new ChildrenProbabilityCollection();
//Act
//Here the mothers are randomly assigned to a designated amount of children.
//The reults are first put in a list and after that they get sorted.
for (int i = 0; i < 32; i++)
{
childAmount.Add(i, new List <double>());
childChoice = new List <double>();
int age = i + 18;
weights = MotherData.GetColumn <double>(Convert.ToString(age)).Values.Select(c => Convert.ToDouble(c)).ToList();
for (int j = 0; j < womenAges[i]; j++)
{
Children.WeightGetChildren = weights;
childChoice.Add(env.RandChoice(Children.SourceGetChildren, Children.WeightGetChildren));
}
var children = childChoice.Where(p => p == 1).Count();
//childAmount = new List<double>();
for (int j = 0; j < children; j++)
{
childAmount[i].Add(env.RandChoice(Children.SourceAmountChildren, Children.WeightsAmountChildren));
}
//Here they are getting sorted and added to a new collection
collectionNew.Add(new ChildrenProbability()
{
Age = age,
childrenAmount = { }
//WomenNoChildren = childChoice.Where(p => p == 0).Count(),
//MotherOfOneAgeFirstBorn = childAmount[i].Where(p => p == 1).Count(),
//MotherOfTwoAgeFirtBorn = childAmount[i].Where(p => p == 2).Count(),
//MotherOfThreeAgeFirstBorn = childAmount[i].Where(p => p == 3).Count(),
//BirthingMothersTotal = children
});
}
//Both collections get exported out
var exportNew = Frame.FromRecords(collectionNew);
var export = Frame.FromRecords(collection);
exportNew.SaveCsv($"{Global.GetDataFolder()}childrenProbabilityDifferent.csv");
//Assert
int expectedFirstChildren34 = 2564;
int actualFirstChildren34 = childAmount[16].Where(p => p == 1).Count();
Assert.Equal(expectedFirstChildren34, actualFirstChildren34);
int expectedSecondChildren18 = 6;
int actualSecondChildren18 = childAmount[0].Where(p => p == 2).Count();
Assert.Equal(expectedSecondChildren18, actualSecondChildren18);
int expectedThirdChildren43 = 73;
int actualThirdChildren43 = childAmount[25].Where(p => p == 3).Count();
Assert.Equal(expectedThirdChildren43, actualThirdChildren43);
}
public void MaritalAgePropability()
{
//Arrange
var env = new SimSharp.ThreadSafeSimulation(42);
Frame <int, string> frame;
frame = Frame.ReadCsv("../../../../../doc/data/MaritalAgeWeights.csv");
List <double> femaleWeights;
List <double> maleWeights;
var womenList = new Dictionary <int, Dictionary <PartnerType, int> >();
var menList = new Dictionary <int, Dictionary <PartnerType, int> >();
var collection = new MaritalAgePropabilityCollection();
var femaleAge = Age.FemaleWeights;
var maleAge = Age.MaleWeights;
//Act
//In here we decide what marital status someone has at a certain age.
//In our case it goes from 18 to 99 years old.
for (int i = 18; i < 100; i++)
{
//We made a dictionary for both men and women.
womenList.Add(i, new Dictionary <PartnerType, int>());
womenList[i].Add(PartnerType.Single, 0);
womenList[i].Add(PartnerType.Married, 0);
womenList[i].Add(PartnerType.Partnership, 0);
menList.Add(i, new Dictionary <PartnerType, int>());
menList[i].Add(PartnerType.Single, 0);
menList[i].Add(PartnerType.Married, 0);
menList[i].Add(PartnerType.Partnership, 0);
maleWeights = frame.GetColumn <double>(Convert.ToString(i)).Values.Select(c => Convert.ToDouble(c)).Take(3).ToList();
femaleWeights = frame.GetColumn <double>(Convert.ToString(i)).Values.Select(c => Convert.ToDouble(c)).Skip(3).Take(3).ToList();
//Female age is all the predictated ages of females in 2020.
//This can be found in the probability class "Age".
//First they get the weights from their age group. And with those weights deciding the results in a randomchoice, they get put into their dictionary.
for (int j = 0; j < femaleAge[i]; j++)
{
MaritalStatus.Weights = femaleWeights;
womenList[i][env.RandChoice(MaritalStatus.Source, MaritalStatus.Weights)]++;
}
//The exact same happens as stated above in women.
for (int j = 0; j < maleAge[i]; j++)
{
MaritalStatus.Weights = maleWeights;
menList[i][env.RandChoice(MaritalStatus.Source, MaritalStatus.Weights)]++;
}
//In here we sort everything into their respective classes.
//i here is yet again representing an age group.
collection.Add(new MaritalAgeProbability()
{
SingleWomen = womenList[i][PartnerType.Single],
MarriedWomen = womenList[i][PartnerType.Married],
DivorcedWomen = womenList[i][PartnerType.Partnership],
SingleMen = womenList[i][PartnerType.Single],
MarriedMen = womenList[i][PartnerType.Married],
DivorcedMen = womenList[i][PartnerType.Partnership],
Age = i
});
}
//Putting the collection into a CSV which can be found in the location stated below.
var export = Frame.FromRecords(collection);
export.SaveCsv($"{Global.GetDataFolder()}maritalAgeProbability.csv");
//Assert
//We asserted a few ages to see if the test would keep on the same results as before.
//For women we used ages 18, 36 and 98 as for men we used 30, 57 and 67.
Assert.Equal(42679, womenList[36][PartnerType.Single]);
Assert.Equal(92, womenList[18][PartnerType.Married]);
Assert.Equal(741, womenList[98][PartnerType.Partnership]);
Assert.Equal(9900, menList[67][PartnerType.Single]);
Assert.Equal(81637, menList[57][PartnerType.Married]);
Assert.Equal(1603, menList[30][PartnerType.Partnership]);
}
