make_sample_data.py

"""Generate the synthetic sample dataset shipped with this Space.

Everything here is fabricated with a fixed seed — no real, customer, or
proprietary data is used anywhere in this project. Re-run to regenerate:

    python make_sample_data.py
"""

import csv
import random
from datetime import date, timedelta
from pathlib import Path

SEED = 20260811
ROWS = 2000
OUT = Path(__file__).parent / "data" / "sample_sales.csv"

REGIONS = {
    "North America": ["United States", "United States", "Canada"],
    "LATAM": ["Mexico", "Mexico", "Colombia", "Argentina", "Chile"],
    "EMEA": ["Spain", "Germany", "United Kingdom"],
}
REGION_WEIGHTS = [0.45, 0.35, 0.20]

CATALOG = {
    "Electronics": [
        ("Wireless Headphones", 89.99),
        ("4K Monitor", 329.00),
        ("Mechanical Keyboard", 119.50),
        ("Smart Speaker", 59.00),
    ],
    "Home & Kitchen": [
        ("Espresso Machine", 249.00),
        ("Air Fryer", 129.99),
        ("Cookware Set", 179.00),
        ("Blender", 74.50),
    ],
    "Apparel": [
        ("Running Jacket", 94.00),
        ("Denim Jeans", 68.00),
        ("Merino Socks", 22.00),
        ("Rain Shell", 139.00),
    ],
    "Sports": [
        ("Yoga Mat", 45.00),
        ("Dumbbell Set", 159.00),
        ("Trail Backpack", 112.00),
        ("Water Bottle", 28.00),
    ],
    "Beauty": [
        ("Skincare Set", 82.00),
        ("Hair Dryer", 96.00),
        ("Perfume", 115.00),
        ("Electric Razor", 68.50),
    ],
}
CATEGORY_WEIGHTS = [0.28, 0.22, 0.20, 0.16, 0.14]

CHANNELS = ["Online", "Retail Store", "Partner"]
CHANNEL_WEIGHTS = [0.58, 0.28, 0.14]

SEGMENTS = ["Consumer", "Small Business", "Enterprise"]
SEGMENT_WEIGHTS = [0.62, 0.26, 0.12]

# Retail seasonality: quiet start, holiday spike in Nov/Dec.
MONTH_WEIGHTS = [0.60, 0.58, 0.70, 0.75, 0.80, 0.80,
                 0.85, 0.80, 0.90, 1.00, 1.60, 1.80]

START = date(2025, 1, 1)
DAYS = 365


def main() -> None:
    rng = random.Random(SEED)
    day_weights = [MONTH_WEIGHTS[(START + timedelta(days=d)).month - 1]
                   for d in range(DAYS)]

    rows = []
    for i in range(ROWS):
        order_date = START + timedelta(
            days=rng.choices(range(DAYS), weights=day_weights)[0]
        )
        region = rng.choices(list(REGIONS), weights=REGION_WEIGHTS)[0]
        country = rng.choice(REGIONS[region])
        category = rng.choices(list(CATALOG), weights=CATEGORY_WEIGHTS)[0]
        product, base_price = rng.choice(CATALOG[category])
        channel = rng.choices(CHANNELS, weights=CHANNEL_WEIGHTS)[0]
        segment = rng.choices(SEGMENTS, weights=SEGMENT_WEIGHTS)[0]

        quantity = rng.choices([1, 2, 3, 4, 6, 10],
                               weights=[46, 24, 12, 8, 6, 4])[0]
        if segment == "Enterprise":
            quantity *= rng.choice([2, 3, 5])

        # Prices drift a little by market; discounts cluster around promos.
        unit_price = round(base_price * rng.uniform(0.94, 1.08), 2)
        discount_pct = rng.choices(
            [0.00, 0.05, 0.10, 0.15, 0.25],
            weights=[52, 18, 15, 9, 6],
        )[0]
        if order_date.month in (11, 12) and rng.random() < 0.45:
            discount_pct = max(discount_pct, 0.15)

        revenue = round(quantity * unit_price * (1 - discount_pct), 2)

        rows.append({
            "order_id": f"ORD-{100000 + i}",
            "order_date": order_date.isoformat(),
            "region": region,
            "country": country,
            "channel": channel,
            "customer_segment": segment,
            "category": category,
            "product": product,
            "quantity": quantity,
            "unit_price": unit_price,
            "discount_pct": discount_pct,
            "revenue": revenue,
        })

    rows.sort(key=lambda r: r["order_date"])

    OUT.parent.mkdir(parents=True, exist_ok=True)
    with OUT.open("w", newline="", encoding="utf-8") as fh:
        writer = csv.DictWriter(fh, fieldnames=list(rows[0]))
        writer.writeheader()
        writer.writerows(rows)

    total = sum(r["revenue"] for r in rows)
    print(f"wrote {len(rows)} rows to {OUT}  (total revenue {total:,.2f})")


if __name__ == "__main__":
    main()