Machine Learning

from datetime import datetime

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

# Set random seed for reproducibility using Generator
rng = np.random.default_rng(42)

# Create 2-year daily time series
start_date = datetime(2022, 1, 1)
end_date = datetime(2023, 12, 31)
dates = pd.date_range(start=start_date, end=end_date, freq="D")
n_days = len(dates)

# Create day of year for seasonal calculations (0-365)
day_of_year = np.array([d.timetuple().tm_yday for d in dates])

# Seasonal parameters
winter_mean = 23.0  # Higher β in winter (warmer indoor preference)
summer_mean = 21.0  # Lower β in summer
winter_std = 1.2  # Lower variability in winter
summer_std = 2.0  # Higher variability in summer

# Create seasonal mean using sinusoidal function
# Peak in winter (day ~15 = mid January), minimum in summer (day ~195 = mid July)
seasonal_mean = summer_mean + (winter_mean - summer_mean) * 0.5 * (
    1 + np.cos(2 * np.pi * (day_of_year - 15) / 365)
)

# Create seasonal standard deviation
# Minimum std in winter, maximum in summer
seasonal_std = winter_std + (summer_std - winter_std) * 0.5 * (
    1 - np.cos(2 * np.pi * (day_of_year - 15) / 365)
)

# Generate β values from Gaussian distributions with seasonal parameters
beta_values = rng.normal(seasonal_mean, seasonal_std)

# Create DataFrame for easier handling
df = pd.DataFrame(
    {
        "date": dates,
        "beta": beta_values,
        "seasonal_mean": seasonal_mean,
        "seasonal_std": seasonal_std,
        "month": [d.month for d in dates],
    }
)

print(f"Simulatie van β-waarden over {len(dates)} dagen (2 jaar)")
print(f"Gemiddelde β: {np.mean(beta_values):.2f}°C")
print(f"Standaarddeviatie β: {np.std(beta_values):.2f}°C")
print(f"Minimum β: {np.min(beta_values):.2f}°C")
print(f"Maximum β: {np.max(beta_values):.2f}°C")

# Create comprehensive visualization
fig, (ax1, ax2, ax3) = plt.subplots(3, 1, figsize=(14, 12))

# Plot 1: β values over time with seasonal mean
ax1.plot(df["date"], df["beta"], alpha=0.6, color="steelblue", linewidth=0.8, label="β waarden")
ax1.plot(df["date"], df["seasonal_mean"], color="red", linewidth=2, label="Seizoensgemiddelde")
ax1.fill_between(
    df["date"],
    df["seasonal_mean"] - df["seasonal_std"],
    df["seasonal_mean"] + df["seasonal_std"],
    alpha=0.2,
    color="red",
    label="±1 standaarddeviatie",
)
ax1.set_ylabel("β (°C)")
ax1.set_title("Simulatie van airco drempelwaarde β over 2 jaar")
ax1.legend()
ax1.grid(True, alpha=0.3)

# Add season labels
season_colors = {
    "Winter": "lightblue",
    "Lente": "lightgreen",
    "Zomer": "yellow",
    "Herfst": "orange",
}
for year in [2022, 2023]:
    ax1.axvspan(
        datetime(year, 12, 21), datetime(year + 1, 3, 20), alpha=0.1, color=season_colors["Winter"]
    )
    ax1.axvspan(
        datetime(year, 3, 20), datetime(year, 6, 21), alpha=0.1, color=season_colors["Lente"]
    )
    ax1.axvspan(
        datetime(year, 6, 21), datetime(year, 9, 22), alpha=0.1, color=season_colors["Zomer"]
    )
    ax1.axvspan(
        datetime(year, 9, 22), datetime(year, 12, 21), alpha=0.1, color=season_colors["Herfst"]
    )

# Plot 2: Monthly boxplot
monthly_data = [df[df["month"] == month]["beta"].values for month in range(1, 13)]
box_plot = ax2.boxplot(
    monthly_data,
    tick_labels=[
        "Jan",
        "Feb",
        "Mrt",
        "Apr",
        "Mei",
        "Jun",
        "Jul",
        "Aug",
        "Sep",
        "Okt",
        "Nov",
        "Dec",
    ],
    patch_artist=True,
)
ax2.set_ylabel("β (°C)")
ax2.set_title("Maandelijkse verdeling van β-waarden")
ax2.grid(True, alpha=0.3)

# Color boxes by season
season_months = {
    "Winter": [0, 1, 11],  # Jan, Feb, Dec (0-indexed)
    "Lente": [2, 3, 4],  # Mar, Apr, May
    "Zomer": [5, 6, 7],  # Jun, Jul, Aug
    "Herfst": [8, 9, 10],  # Sep, Oct, Nov
}

for season, months in season_months.items():
    for month_idx in months:
        box_plot["boxes"][month_idx].set_facecolor(season_colors[season])

# Plot 3: Seasonal statistics
seasonal_stats = (
    df.groupby(df["date"].dt.month)
    .agg({"beta": ["mean", "std"], "seasonal_mean": "mean", "seasonal_std": "mean"})
    .round(2)
)

months = range(1, 13)
ax3.plot(
    months,
    seasonal_stats[("beta", "mean")],
    "o-",
    color="blue",
    label="Werkelijke gemiddelde",
    linewidth=2,
)
ax3.plot(
    months,
    seasonal_stats[("seasonal_mean", "mean")],
    "s-",
    color="red",
    label="Theoretische gemiddelde",
    linewidth=2,
)
ax3.set_ylabel("Gemiddelde β (°C)")
ax3.set_xlabel("Maand")
ax3.set_title("Seizoenspatroon in gemiddelde β-waarden")
ax3.set_xticks(months)
ax3.set_xticklabels(
    ["Jan", "Feb", "Mrt", "Apr", "Mei", "Jun", "Jul", "Aug", "Sep", "Okt", "Nov", "Dec"]
)
ax3.legend()
ax3.grid(True, alpha=0.3)

plt.tight_layout()
plt.show()

# Print seasonal summary
print("\nSeizoensstatistieken:")
season_mapping = {
    "Winter": [12, 1, 2],
    "Lente": [3, 4, 5],
    "Zomer": [6, 7, 8],
    "Herfst": [9, 10, 11],
}

for season, months in season_mapping.items():
    season_data = df[df["month"].isin(months)]["beta"]
    print(f"{season}: Gemiddelde = {season_data.mean():.2f}°C, Std = {season_data.std():.2f}°C")