Weather Simulation and Analysis Pipeline

import numpy as np
import pandas as pd

from pipefunc import Pipeline, pipefunc


# Step 1: Simulate Temperature Data
@pipefunc(output_name="temperature", mapspec="city[i], day[j] -> temperature[i, j]")
def simulate_temperature(city: str, day) -> float:
    np.random.seed(hash(city) % 2**32)  # For reproducibility
    mean_temp = 20 + (hash(city) % 10)  # Base temp varies by city
    temp_variation = 5 * np.sin(day.dayofyear * (2 * np.pi / 365))  # Seasonal variation
    noise = np.random.normal(0, 2)  # Random daily fluctuation
    return float(mean_temp + temp_variation + noise)  # Ensure this is a float


# Step 2: Compute Statistics
@pipefunc(
    output_name=("mean_temp", "variance"),
    mapspec="temperature[i, :] -> mean_temp[i], variance[i]",
    output_picker=dict.__getitem__,
)
def compute_statistics(temperature):
    temp_array = np.array(temperature, dtype=float)  # Ensure it's a numeric array
    mean_temp = np.mean(temp_array)
    var_temp = np.var(temp_array)
    return {"mean_temp": mean_temp, "variance": var_temp}


# Create the pipeline
pipeline_weather = Pipeline([simulate_temperature, compute_statistics])

# Define cities and days
cities = ["New York", "Los Angeles", "Chicago"]
days = pd.date_range("2023-01-01", "2023-01-10")  # 10 days

# Run the pipeline
results = pipeline_weather.map({"city": cities, "day": days}, run_folder="weather_simulation_results")

# Load and display the xarray dataset
weather_dataset = results.to_xarray()
display(weather_dataset)

# Plot temperatures for each city
weather_dataset.temperature.plot.line(
    x="day",
    hue="city",
    marker="o",
    figsize=(12, 6),
)
Weather Simulation and Analysis Pipeline#
