{
  "cells": [
    {
      "cell_type": "markdown",
      "id": "9902bac1",
      "metadata": {},
      "source": [
        "# AdventureWorks Sales Forecast Analysis\n",
        "\n",
        "This notebook integrates the complete analysis workflow from raw AdventureWorks sales data to forecast simulations and segment-level volatility analysis. It is designed to run from top to bottom and regenerate the CSV outputs and figures used in the report.\n",
        "\n",
        "Workflow:\n",
        "\n",
        "1. Data preparation and star-schema joins\n",
        "2. Part 1 Monte Carlo forecast using TTM YoY growth\n",
        "3. Part 2 daily bootstrap forecast\n",
        "4. Model comparison\n",
        "5. Category and Channel extended bootstrap analysis"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "a32da56f",
      "metadata": {},
      "source": [
        "## Setup\n",
        "\n",
        "This section imports packages, defines project paths, and creates output folders. The source workbook is expected at `work/AdventureWorks Sales.xlsx`. All generated CSV files are saved under `outputs/csv`; all generated figures are saved under `outputs/figures`."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "id": "6232fe96",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Project root: C:\\Users\\ZhuanZ（无密码）\\Documents\\Codex\\2026-07-02\\codex-spec-codex-codex-adventureworks-forecast\n",
            "Source workbook: C:\\Users\\ZhuanZ（无密码）\\Documents\\Codex\\2026-07-02\\codex-spec-codex-codex-adventureworks-forecast\\work\\AdventureWorks Sales.xlsx\n",
            "Prepared CSV fallback available: True\n"
          ]
        }
      ],
      "source": [
        "from pathlib import Path\n",
        "import warnings\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "warnings.filterwarnings(\"ignore\", category=FutureWarning)\n",
        "\n",
        "# Resolve paths robustly. In desktop Jupyter, this will usually be the project root.\n",
        "# In JupyterLite/Pyodide, Path.cwd() is the browser virtual filesystem, usually /drive.\n",
        "PROJECT_ROOT = Path.cwd()\n",
        "for parent in [PROJECT_ROOT, *PROJECT_ROOT.parents]:\n",
        "    if (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "        PROJECT_ROOT = parent\n",
        "        break\n",
        "    if (parent / \"outputs\" / \"csv\" / \"daily_sales.csv\").exists():\n",
        "        PROJECT_ROOT = parent\n",
        "        break\n",
        "\n",
        "DATA_CANDIDATES = [\n",
        "    PROJECT_ROOT / \"work\" / \"AdventureWorks Sales.xlsx\",\n",
        "    PROJECT_ROOT / \"AdventureWorks Sales.xlsx\",\n",
        "    Path.cwd() / \"work\" / \"AdventureWorks Sales.xlsx\",\n",
        "    Path.cwd() / \"AdventureWorks Sales.xlsx\",\n",
        "]\n",
        "DATA_PATH = next((p for p in DATA_CANDIDATES if p.exists()), None)\n",
        "\n",
        "CSV_DIR = PROJECT_ROOT / \"outputs\" / \"csv\"\n",
        "FIG_DIR = PROJECT_ROOT / \"outputs\" / \"figures\"\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "PREPARED_CSV_FILES = [\n",
        "    CSV_DIR / \"daily_sales.csv\",\n",
        "    CSV_DIR / \"annual_sales.csv\",\n",
        "    CSV_DIR / \"daily_category_sales.csv\",\n",
        "    CSV_DIR / \"daily_channel_sales.csv\",\n",
        "]\n",
        "PREPARED_CSV_AVAILABLE = all(p.exists() for p in PREPARED_CSV_FILES)\n",
        "\n",
        "pd.set_option(\"display.float_format\", lambda x: f\"{x:,.4f}\")\n",
        "print(f\"Project root: {PROJECT_ROOT}\")\n",
        "print(f\"Source workbook: {DATA_PATH if DATA_PATH else 'not found'}\")\n",
        "print(f\"Prepared CSV fallback available: {PREPARED_CSV_AVAILABLE}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "fdbe7558",
      "metadata": {},
      "source": [
        "## Helper Functions\n",
        "\n",
        "The notebook uses a small set of helpers for date parsing, complete-year checks, summaries, and display formatting. All stochastic simulation uses NumPy's standard `np.random.default_rng` interface with fixed seeds for reproducibility."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "id": "ae3d7cb6",
      "metadata": {},
      "outputs": [],
      "source": [
        "def parse_order_date_key(series: pd.Series) -> pd.Series:\n",
        "    keys = pd.to_numeric(series, errors=\"coerce\")\n",
        "    parsed = pd.Series(pd.NaT, index=series.index, dtype=\"datetime64[ns]\")\n",
        "    valid = keys.notna()\n",
        "    parsed.loc[valid] = pd.to_datetime(\n",
        "        keys.loc[valid].astype(\"int64\").astype(str),\n",
        "        format=\"%Y%m%d\",\n",
        "        errors=\"coerce\",\n",
        "    )\n",
        "    return parsed\n",
        "\n",
        "\n",
        "def complete_calendar_years(min_date: pd.Timestamp, max_date: pd.Timestamp) -> list[int]:\n",
        "    years = []\n",
        "    for year in range(min_date.year, max_date.year + 1):\n",
        "        if min_date <= pd.Timestamp(year=year, month=1, day=1) and max_date >= pd.Timestamp(year=year, month=12, day=31):\n",
        "            years.append(year)\n",
        "    return years\n",
        "\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "\n",
        "def pct(x):\n",
        "    return f\"{x:.2%}\""
      ]
    },
    {
      "cell_type": "markdown",
      "id": "96cc53fc",
      "metadata": {},
      "source": [
        "## Step 1: Data Preparation and Star-Schema Join\n",
        "\n",
        "This step reads all seven workbook sheets, joins the Sales fact table to the dimension tables, converts `OrderDateKey` into a date, and constructs daily and annual sales series. The daily series is reindexed to a complete calendar so that missing dates would be represented as zero sales rather than silently omitted. Annual sales keeps only complete calendar years to avoid distorted YoY growth from partial years."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "id": "2a284a12",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Workbook sheets:\n",
            "- Sales Order_data: 121,253 rows, 4 columns\n",
            "- Sales Territory_data: 11 rows, 4 columns\n",
            "- Sales_data: 121,253 rows, 15 columns\n",
            "- Reseller_data: 702 rows, 8 columns\n",
            "- Date_data: 1,461 rows, 7 columns\n",
            "- Product_data: 397 rows, 9 columns\n",
            "- Customer_data: 18,485 rows, 7 columns\n"
          ]
        }
      ],
      "source": [
        "# Read workbook sheets when the Excel file is available.\n",
        "# If running in JupyterLite and the workbook has not been uploaded, the next cell can fall back\n",
        "# to prepared CSV files under outputs/csv.\n",
        "sheets = None\n",
        "excel_read_error = None\n",
        "if DATA_PATH is not None:\n",
        "    try:\n",
        "        sheets = pd.read_excel(DATA_PATH, sheet_name=None)\n",
        "        print(\"Workbook sheets:\")\n",
        "        for name, frame in sheets.items():\n",
        "            print(f\"- {name}: {frame.shape[0]:,} rows, {frame.shape[1]:,} columns\")\n",
        "    except Exception as exc:\n",
        "        excel_read_error = exc\n",
        "        print(\"Excel file was found but could not be read in this environment.\")\n",
        "        print(type(exc).__name__ + \": \" + str(exc))\n",
        "        print(\"If this is JupyterLite/Pyodide, upload the prepared outputs/csv files or run the notebook in regular Jupyter with openpyxl installed.\")\n",
        "else:\n",
        "    print(\"Workbook not found in this Jupyter filesystem.\")\n",
        "    print(\"In JupyterLite, upload AdventureWorks Sales.xlsx to /drive/work/ or upload the prepared outputs/csv folder.\")\n",
        "\n",
        "if sheets is not None:\n",
        "    sales_raw = sheets[\"Sales_data\"].copy()\n",
        "    sales_order = sheets[\"Sales Order_data\"].copy()\n",
        "    territory = sheets[\"Sales Territory_data\"].copy()\n",
        "    reseller = sheets[\"Reseller_data\"].copy()\n",
        "    date_dim = sheets[\"Date_data\"].copy()\n",
        "    product = sheets[\"Product_data\"].copy()\n",
        "    customer = sheets[\"Customer_data\"].copy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "id": "87fa700d",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Step 1 complete\n",
            "                          check      value\n",
            "              sales_source_rows     121253\n",
            "             joined_output_rows     121253\n",
            "                       date_min 2017-07-01\n",
            "                       date_max 2020-06-15\n",
            "    calendar_days_after_reindex       1081\n",
            "  zero_sales_days_after_reindex          0\n",
            "   complete_calendar_years_used  2018;2019\n",
            "    invalid_order_dates_dropped          0\n",
            "      missing_sales_amount_rows          0\n",
            "     negative_sales_amount_rows          0\n",
            "         zero_sales_amount_rows          0\n",
            "duplicate_sales_order_line_keys          0\n",
            "     unmatched_sales_order_rows          0\n",
            "       unmatched_territory_rows          0\n",
            "        unmatched_reseller_rows          0\n",
            "        unmatched_customer_rows          0\n",
            "         unmatched_product_rows          0\n",
            "            unmatched_date_rows          0\n",
            "\n",
            "Annual sales used for complete-year modeling:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>year</th>\n",
              "      <th>sales_amount</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2018</td>\n",
              "      <td>30,516,891.7954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2019</td>\n",
              "      <td>42,895,109.5011</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   year    sales_amount\n",
              "0  2018 30,516,891.7954\n",
              "1  2019 42,895,109.5011"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Daily sales describe:\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "count     1,081.0000\n",
              "mean    101,581.1972\n",
              "std      77,437.6180\n",
              "min       4,009.3300\n",
              "25%      39,821.9000\n",
              "50%      84,324.9841\n",
              "75%     143,032.2955\n",
              "max     569,749.7212\n",
              "Name: sales_amount, dtype: float64"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "if sheets is not None:\n",
        "    # Clean dates and numeric sales values.\n",
        "    sales = sales_raw.copy()\n",
        "    sales[\"date\"] = parse_order_date_key(sales[\"OrderDateKey\"])\n",
        "    invalid_dates = int(sales[\"date\"].isna().sum())\n",
        "    sales = sales.dropna(subset=[\"date\"]).copy()\n",
        "    sales[\"date\"] = sales[\"date\"].dt.normalize()\n",
        "    sales[\"Sales Amount\"] = pd.to_numeric(sales[\"Sales Amount\"], errors=\"coerce\")\n",
        "\n",
        "    # Join dimensions.\n",
        "    joined = sales.merge(\n",
        "        sales_order[[\"SalesOrderLineKey\", \"Channel\", \"Sales Order\", \"Sales Order Line\"]],\n",
        "        on=\"SalesOrderLineKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "    joined = joined.merge(\n",
        "        product[[\"ProductKey\", \"SKU\", \"Product\", \"Color\", \"Model\", \"Subcategory\", \"Category\"]],\n",
        "        on=\"ProductKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "    joined = joined.merge(\n",
        "        territory[[\"SalesTerritoryKey\", \"Region\", \"Country\", \"Group\"]].rename(\n",
        "            columns={\"Country\": \"Territory Country\", \"Group\": \"Territory Group\"}\n",
        "        ),\n",
        "        on=\"SalesTerritoryKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "    joined = joined.merge(\n",
        "        reseller[[\"ResellerKey\", \"Business Type\", \"Reseller\", \"City\", \"State-Province\", \"Country-Region\"]].rename(\n",
        "            columns={\n",
        "                \"City\": \"Reseller City\",\n",
        "                \"State-Province\": \"Reseller State-Province\",\n",
        "                \"Country-Region\": \"Reseller Country-Region\",\n",
        "            }\n",
        "        ),\n",
        "        on=\"ResellerKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "    joined = joined.merge(\n",
        "        customer[[\"CustomerKey\", \"Customer\", \"City\", \"State-Province\", \"Country-Region\"]].rename(\n",
        "            columns={\n",
        "                \"City\": \"Customer City\",\n",
        "                \"State-Province\": \"Customer State-Province\",\n",
        "                \"Country-Region\": \"Customer Country-Region\",\n",
        "            }\n",
        "        ),\n",
        "        on=\"CustomerKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "    joined = joined.merge(\n",
        "        date_dim[[\"DateKey\", \"Fiscal Year\", \"Fiscal Quarter\", \"Month\", \"Full Date\", \"MonthKey\"]],\n",
        "        left_on=\"OrderDateKey\",\n",
        "        right_on=\"DateKey\",\n",
        "        how=\"left\",\n",
        "    )\n",
        "\n",
        "    joined[\"Category\"] = joined[\"Category\"].fillna(\"Unknown\")\n",
        "    joined[\"Channel\"] = joined[\"Channel\"].fillna(\"Unknown\")\n",
        "    joined[\"Sales Amount\"] = joined[\"Sales Amount\"].fillna(0.0)\n",
        "\n",
        "    min_date = joined[\"date\"].min()\n",
        "    max_date = joined[\"date\"].max()\n",
        "    calendar = pd.date_range(min_date, max_date, freq=\"D\")\n",
        "\n",
        "    daily_sales = (\n",
        "        joined.groupby(\"date\", as_index=True)[\"Sales Amount\"]\n",
        "        .sum()\n",
        "        .reindex(calendar, fill_value=0.0)\n",
        "        .rename_axis(\"date\")\n",
        "        .reset_index(name=\"sales_amount\")\n",
        "    )\n",
        "\n",
        "    complete_years = complete_calendar_years(min_date, max_date)\n",
        "    annual_all = daily_sales.assign(year=daily_sales[\"date\"].dt.year).groupby(\"year\", as_index=False)[\"sales_amount\"].sum()\n",
        "    annual_sales = annual_all[annual_all[\"year\"].isin(complete_years)].reset_index(drop=True)\n",
        "\n",
        "    category_index = pd.MultiIndex.from_product([calendar, sorted(joined[\"Category\"].dropna().unique())], names=[\"date\", \"category\"])\n",
        "    daily_category_sales = (\n",
        "        joined.groupby([\"date\", \"Category\"], as_index=True)[\"Sales Amount\"]\n",
        "        .sum()\n",
        "        .rename_axis([\"date\", \"category\"])\n",
        "        .reindex(category_index, fill_value=0.0)\n",
        "        .reset_index(name=\"sales_amount\")\n",
        "    )\n",
        "\n",
        "    channel_index = pd.MultiIndex.from_product([calendar, sorted(joined[\"Channel\"].dropna().unique())], names=[\"date\", \"channel\"])\n",
        "    daily_channel_sales = (\n",
        "        joined.groupby([\"date\", \"Channel\"], as_index=True)[\"Sales Amount\"]\n",
        "        .sum()\n",
        "        .rename_axis([\"date\", \"channel\"])\n",
        "        .reindex(channel_index, fill_value=0.0)\n",
        "        .reset_index(name=\"sales_amount\")\n",
        "    )\n",
        "\n",
        "    data_quality = pd.DataFrame([\n",
        "        (\"sales_source_rows\", len(sales_raw)),\n",
        "        (\"joined_output_rows\", len(joined)),\n",
        "        (\"date_min\", min_date.date().isoformat()),\n",
        "        (\"date_max\", max_date.date().isoformat()),\n",
        "        (\"calendar_days_after_reindex\", len(daily_sales)),\n",
        "        (\"zero_sales_days_after_reindex\", int((daily_sales[\"sales_amount\"] == 0).sum())),\n",
        "        (\"complete_calendar_years_used\", \";\".join(map(str, complete_years))),\n",
        "        (\"invalid_order_dates_dropped\", invalid_dates),\n",
        "        (\"missing_sales_amount_rows\", int(pd.to_numeric(sales_raw[\"Sales Amount\"], errors=\"coerce\").isna().sum())),\n",
        "        (\"negative_sales_amount_rows\", int((pd.to_numeric(sales_raw[\"Sales Amount\"], errors=\"coerce\") < 0).sum())),\n",
        "        (\"zero_sales_amount_rows\", int((pd.to_numeric(sales_raw[\"Sales Amount\"], errors=\"coerce\") == 0).sum())),\n",
        "        (\"duplicate_sales_order_line_keys\", int(sales_raw[\"SalesOrderLineKey\"].duplicated().sum())),\n",
        "        (\"unmatched_sales_order_rows\", int(joined[\"Channel\"].isna().sum())),\n",
        "        (\"unmatched_territory_rows\", int(joined[\"Region\"].isna().sum())),\n",
        "        (\"unmatched_reseller_rows\", int(joined[\"Reseller\"].isna().sum())),\n",
        "        (\"unmatched_customer_rows\", int(joined[\"Customer\"].isna().sum())),\n",
        "        (\"unmatched_product_rows\", int(joined[\"Product\"].isna().sum())),\n",
        "        (\"unmatched_date_rows\", int(joined[\"DateKey\"].isna().sum())),\n",
        "    ], columns=[\"check\", \"value\"])\n",
        "\n",
        "    daily_desc = daily_sales[\"sales_amount\"].describe()\n",
        "    for idx, val in daily_desc.items():\n",
        "        data_quality.loc[len(data_quality)] = [f\"daily_sales_describe_{idx}\", val]\n",
        "\n",
        "    joined.to_csv(CSV_DIR / \"joined_sales_step1.csv\", index=False)\n",
        "    daily_sales.to_csv(CSV_DIR / \"daily_sales.csv\", index=False)\n",
        "    annual_sales.to_csv(CSV_DIR / \"annual_sales.csv\", index=False)\n",
        "    daily_category_sales.to_csv(CSV_DIR / \"daily_category_sales.csv\", index=False)\n",
        "    daily_channel_sales.to_csv(CSV_DIR / \"daily_channel_sales.csv\", index=False)\n",
        "    data_quality.to_csv(CSV_DIR / \"data_quality_step1.csv\", index=False)\n",
        "else:\n",
        "    required = {p.name: p for p in PREPARED_CSV_FILES}\n",
        "    missing = [name for name, p in required.items() if not p.exists()]\n",
        "    if missing:\n",
        "        raise FileNotFoundError(\n",
        "            \"Neither the source Excel workbook nor the prepared CSV fallback files are available in this Jupyter filesystem. \"\n",
        "            \"In JupyterLite, upload either work/AdventureWorks Sales.xlsx or the outputs/csv folder. \"\n",
        "            f\"Missing fallback CSVs: {missing}. CSV folder checked: {CSV_DIR}\"\n",
        "        )\n",
        "    print(\"Using prepared CSV fallback files from outputs/csv. Step 1 raw Excel join is skipped in this JupyterLite run.\")\n",
        "    joined = None\n",
        "    daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "    data_quality_path = CSV_DIR / \"data_quality_step1.csv\"\n",
        "    data_quality = pd.read_csv(data_quality_path) if data_quality_path.exists() else pd.DataFrame(columns=[\"check\", \"value\"])\n",
        "    min_date = daily_sales[\"date\"].min()\n",
        "    max_date = daily_sales[\"date\"].max()\n",
        "    calendar = pd.date_range(min_date, max_date, freq=\"D\")\n",
        "    complete_years = annual_sales[\"year\"].astype(int).tolist()\n",
        "\n",
        "print(\"Step 1 complete\")\n",
        "if len(data_quality):\n",
        "    print(data_quality.head(18).to_string(index=False))\n",
        "print(\"\\nAnnual sales used for complete-year modeling:\")\n",
        "display(annual_sales)\n",
        "print(\"\\nDaily sales describe:\")\n",
        "display(daily_sales[\"sales_amount\"].describe())"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "86344ea3",
      "metadata": {},
      "source": [
        "## Part 1: Monte Carlo Forecast Using TTM YoY Growth\n",
        "\n",
        "The raw data has only two complete calendar years, 2018 and 2019. That gives only one annual YoY growth rate, which is insufficient for estimating a meaningful standard deviation. To avoid inventing a false sigma, this notebook estimates growth parameters from month-end Trailing Twelve Months (TTM) YoY growth.\n",
        "\n",
        "For each month end, TTM sales is the prior 12 months of sales. TTM YoY compares that value with the TTM value one year earlier. This produces 12 growth observations from 2019-06-30 through 2020-05-31. This is a rigorous adjustment for short history, but the overlapping windows are not fully independent and the sequence shows a rising trend, so the resulting probability estimates should be interpreted conditionally rather than as unconditional certainty."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "id": "80ec5e41",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "TTM growth sample count: 12\n",
            "mu: 0.4322613770 (43.23%)\n",
            "sigma: 0.0877635375 (8.78%)\n",
            "2018 to 2019 annual YoY cross-check: 0.4056185600 (40.56%)\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>month_end</th>\n",
              "      <th>ttm_sales</th>\n",
              "      <th>prior_year_ttm_sales</th>\n",
              "      <th>ttm_growth</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>2019-06-30</td>\n",
              "      <td>34,070,108.5001</td>\n",
              "      <td>23,860,891.1663</td>\n",
              "      <td>0.4279</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>2019-07-31</td>\n",
              "      <td>34,643,481.5532</td>\n",
              "      <td>25,377,224.8397</td>\n",
              "      <td>0.3651</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>2019-08-31</td>\n",
              "      <td>35,925,845.6797</td>\n",
              "      <td>27,284,123.5897</td>\n",
              "      <td>0.3167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>2019-09-30</td>\n",
              "      <td>37,742,327.6736</td>\n",
              "      <td>28,047,781.9756</td>\n",
              "      <td>0.3456</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>2019-10-31</td>\n",
              "      <td>39,127,189.8510</td>\n",
              "      <td>29,643,387.8989</td>\n",
              "      <td>0.3199</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>2019-11-30</td>\n",
              "      <td>40,667,994.2059</td>\n",
              "      <td>28,489,559.9656</td>\n",
              "      <td>0.4275</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>42,895,109.5011</td>\n",
              "      <td>30,516,891.7954</td>\n",
              "      <td>0.4056</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>2020-01-31</td>\n",
              "      <td>44,282,604.8234</td>\n",
              "      <td>31,036,909.0780</td>\n",
              "      <td>0.4268</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>2020-02-29</td>\n",
              "      <td>45,523,289.0765</td>\n",
              "      <td>29,929,807.4130</td>\n",
              "      <td>0.5210</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>2020-03-31</td>\n",
              "      <td>47,860,687.2524</td>\n",
              "      <td>31,321,368.5814</td>\n",
              "      <td>0.5281</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>2020-04-30</td>\n",
              "      <td>49,457,330.6132</td>\n",
              "      <td>32,208,064.3311</td>\n",
              "      <td>0.5356</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>2020-05-31</td>\n",
              "      <td>51,263,683.2591</td>\n",
              "      <td>32,707,130.2773</td>\n",
              "      <td>0.5674</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    month_end       ttm_sales  prior_year_ttm_sales  ttm_growth\n",
              "12 2019-06-30 34,070,108.5001       23,860,891.1663      0.4279\n",
              "13 2019-07-31 34,643,481.5532       25,377,224.8397      0.3651\n",
              "14 2019-08-31 35,925,845.6797       27,284,123.5897      0.3167\n",
              "15 2019-09-30 37,742,327.6736       28,047,781.9756      0.3456\n",
              "16 2019-10-31 39,127,189.8510       29,643,387.8989      0.3199\n",
              "17 2019-11-30 40,667,994.2059       28,489,559.9656      0.4275\n",
              "18 2019-12-31 42,895,109.5011       30,516,891.7954      0.4056\n",
              "19 2020-01-31 44,282,604.8234       31,036,909.0780      0.4268\n",
              "20 2020-02-29 45,523,289.0765       29,929,807.4130      0.5210\n",
              "21 2020-03-31 47,860,687.2524       31,321,368.5814      0.5281\n",
              "22 2020-04-30 49,457,330.6132       32,208,064.3311      0.5356\n",
              "23 2020-05-31 51,263,683.2591       32,707,130.2773      0.5674"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "# Build month-end TTM series and TTM YoY growth.\n",
        "daily_series = daily_sales.set_index(\"date\")[\"sales_amount\"].sort_index()\n",
        "last_complete_month_end = (daily_series.index.max().to_period(\"M\") - 1).to_timestamp(\"M\")\n",
        "month_ends = pd.date_range(\n",
        "    (daily_series.index.min() + pd.Timedelta(days=364)).to_period(\"M\").to_timestamp(\"M\"),\n",
        "    last_complete_month_end,\n",
        "    freq=\"ME\",\n",
        ")\n",
        "\n",
        "ttm_rows = []\n",
        "for end_date in month_ends:\n",
        "    prior_year_end = end_date - pd.DateOffset(years=1)\n",
        "    start_date = prior_year_end + pd.Timedelta(days=1)\n",
        "    ttm_sales_value = daily_series.loc[start_date:end_date].sum()\n",
        "    ttm_rows.append({\"month_end\": end_date, \"window_start\": start_date, \"window_end\": end_date, \"ttm_sales\": ttm_sales_value})\n",
        "\n",
        "ttm_sales = pd.DataFrame(ttm_rows)\n",
        "ttm_sales[\"prior_year_month_end\"] = ttm_sales[\"month_end\"] - pd.DateOffset(years=1)\n",
        "prior_lookup = ttm_sales.set_index(\"month_end\")[\"ttm_sales\"]\n",
        "ttm_sales[\"prior_year_ttm_sales\"] = ttm_sales[\"prior_year_month_end\"].map(prior_lookup)\n",
        "ttm_growth = ttm_sales.dropna(subset=[\"prior_year_ttm_sales\"]).copy()\n",
        "ttm_growth[\"ttm_growth\"] = (ttm_growth[\"ttm_sales\"] - ttm_growth[\"prior_year_ttm_sales\"]) / ttm_growth[\"prior_year_ttm_sales\"]\n",
        "\n",
        "mu = ttm_growth[\"ttm_growth\"].mean()\n",
        "sigma = ttm_growth[\"ttm_growth\"].std(ddof=1)\n",
        "annual_yoy_2018_to_2019 = (\n",
        "    annual_sales.loc[annual_sales[\"year\"] == 2019, \"sales_amount\"].iloc[0]\n",
        "    - annual_sales.loc[annual_sales[\"year\"] == 2018, \"sales_amount\"].iloc[0]\n",
        ") / annual_sales.loc[annual_sales[\"year\"] == 2018, \"sales_amount\"].iloc[0]\n",
        "\n",
        "ttm_sales.to_csv(CSV_DIR / \"ttm_sales.csv\", index=False)\n",
        "ttm_growth[[\"month_end\", \"ttm_sales\", \"prior_year_ttm_sales\", \"ttm_growth\"]].to_csv(CSV_DIR / \"ttm_growth.csv\", index=False)\n",
        "\n",
        "print(f\"TTM growth sample count: {len(ttm_growth)}\")\n",
        "print(f\"mu: {mu:.10f} ({mu:.2%})\")\n",
        "print(f\"sigma: {sigma:.10f} ({sigma:.2%})\")\n",
        "print(f\"2018 to 2019 annual YoY cross-check: {annual_yoy_2018_to_2019:.10f} ({annual_yoy_2018_to_2019:.2%})\")\n",
        "display(ttm_growth[[\"month_end\", \"ttm_sales\", \"prior_year_ttm_sales\", \"ttm_growth\"]])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "id": "aeedefe4",
      "metadata": {},
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Monte Carlo summary:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>method</th>\n",
              "      <td>TTM YoY Monte Carlo</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>seed</th>\n",
              "      <td>42</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>samples</th>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mu</th>\n",
              "      <td>0.4323</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>sigma</th>\n",
              "      <td>0.0878</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>annual_yoy_2018_to_2019</th>\n",
              "      <td>0.4056</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>baseline_year</th>\n",
              "      <td>2019</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>baseline_sales</th>\n",
              "      <td>42,895,109.5011</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>target_growth</th>\n",
              "      <td>0.1000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>target_sales</th>\n",
              "      <td>47,184,620.4512</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean_forecast</th>\n",
              "      <td>61,398,421.6488</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>p5</th>\n",
              "      <td>55,134,671.8075</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>p95</th>\n",
              "      <td>67,621,311.2077</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>target_probability</th>\n",
              "      <td>0.9999</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                           0\n",
              "method                   TTM YoY Monte Carlo\n",
              "seed                                      42\n",
              "samples                                   12\n",
              "mu                                    0.4323\n",
              "sigma                                 0.0878\n",
              "annual_yoy_2018_to_2019               0.4056\n",
              "baseline_year                           2019\n",
              "baseline_sales               42,895,109.5011\n",
              "target_growth                         0.1000\n",
              "target_sales                 47,184,620.4512\n",
              "mean_forecast                61,398,421.6488\n",
              "p5                           55,134,671.8075\n",
              "p95                          67,621,311.2077\n",
              "target_probability                    0.9999"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "# Monte Carlo simulation.\n",
        "n_trials = 10_000\n",
        "mc_seed = 42\n",
        "baseline_year = 2019\n",
        "baseline_sales = annual_sales.loc[annual_sales[\"year\"] == baseline_year, \"sales_amount\"].iloc[0]\n",
        "target_growth = 0.10\n",
        "target_sales = baseline_sales * (1 + target_growth)\n",
        "\n",
        "rng = np.random.default_rng(mc_seed)\n",
        "sim_growth = rng.normal(mu, sigma, n_trials)\n",
        "sim_sales = baseline_sales * (1 + sim_growth)\n",
        "\n",
        "mc_simulation = pd.DataFrame({\n",
        "    \"trial\": np.arange(1, n_trials + 1),\n",
        "    \"simulated_growth\": sim_growth,\n",
        "    \"simulated_sales\": sim_sales,\n",
        "})\n",
        "mc_summary = pd.DataFrame([{\n",
        "    \"method\": \"TTM YoY Monte Carlo\",\n",
        "    \"seed\": mc_seed,\n",
        "    \"samples\": len(ttm_growth),\n",
        "    \"mu\": mu,\n",
        "    \"sigma\": sigma,\n",
        "    \"annual_yoy_2018_to_2019\": annual_yoy_2018_to_2019,\n",
        "    \"baseline_year\": baseline_year,\n",
        "    \"baseline_sales\": baseline_sales,\n",
        "    \"target_growth\": target_growth,\n",
        "    \"target_sales\": target_sales,\n",
        "    \"mean_forecast\": sim_sales.mean(),\n",
        "    \"p5\": np.percentile(sim_sales, 5),\n",
        "    \"p95\": np.percentile(sim_sales, 95),\n",
        "    \"target_probability\": (sim_sales > target_sales).mean(),\n",
        "}])\n",
        "\n",
        "mc_simulation.to_csv(CSV_DIR / \"mc_simulation.csv\", index=False)\n",
        "mc_summary.to_csv(CSV_DIR / \"mc_summary.csv\", index=False)\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10, 6))\n",
        "ax.hist(sim_sales, bins=50, color=\"#2f6f73\", alpha=0.82, edgecolor=\"white\")\n",
        "ax.axvline(mc_summary.loc[0, \"mean_forecast\"], color=\"#111827\", linewidth=2, label=\"Mean\")\n",
        "ax.axvline(mc_summary.loc[0, \"p5\"], color=\"#b91c1c\", linestyle=\"--\", label=\"P5 / P95\")\n",
        "ax.axvline(mc_summary.loc[0, \"p95\"], color=\"#b91c1c\", linestyle=\"--\")\n",
        "ax.axvline(target_sales, color=\"#7c3aed\", linestyle=\":\", linewidth=2, label=\"10% target\")\n",
        "ax.set_title(\"Monte Carlo Annual Sales Forecast\")\n",
        "ax.set_xlabel(\"Simulated annual sales\")\n",
        "ax.set_ylabel(\"Frequency\")\n",
        "ax.legend()\n",
        "fig.tight_layout()\n",
        "fig.savefig(FIG_DIR / \"mc_histogram.png\", dpi=160)\n",
        "plt.show()\n",
        "\n",
        "print(\"Monte Carlo summary:\")\n",
        "display(mc_summary.T)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "7e5b6aa9",
      "metadata": {},
      "source": [
        "## Part 2: Daily Bootstrap Forecast\n",
        "\n",
        "The bootstrap model samples directly from the empirical daily sales distribution. It does not assume Normal growth and does not extrapolate the recent TTM acceleration. Each simulation draws 365 daily sales values with replacement from the 1,081-day historical pool, then sums the simulated year. This preserves the observed daily volatility level, though it samples days independently and therefore does not enforce calendar sequencing."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "id": "d39761d5",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Sampling pool size: 1081\n",
            "Annual bootstrap summary:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>method</th>\n",
              "      <td>Daily Bootstrap</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>seed</th>\n",
              "      <td>43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pool_size</th>\n",
              "      <td>1081</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>n_sims</th>\n",
              "      <td>1000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>horizon_days</th>\n",
              "      <td>365</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean_forecast</th>\n",
              "      <td>37,103,960.4518</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>p5</th>\n",
              "      <td>34,633,619.5663</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>p95</th>\n",
              "      <td>39,574,655.3738</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                             0\n",
              "method         Daily Bootstrap\n",
              "seed                        43\n",
              "pool_size                 1081\n",
              "n_sims                    1000\n",
              "horizon_days               365\n",
              "mean_forecast  37,103,960.4518\n",
              "p5             34,633,619.5663\n",
              "p95            39,574,655.3738"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Checkpoint daily statistics:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>day</th>\n",
              "      <th>date</th>\n",
              "      <th>mean</th>\n",
              "      <th>median</th>\n",
              "      <th>p5</th>\n",
              "      <th>p95</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>2020-06-16</td>\n",
              "      <td>102,874.0184</td>\n",
              "      <td>83,968.4539</td>\n",
              "      <td>14,133.5500</td>\n",
              "      <td>253,965.0897</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>179</th>\n",
              "      <td>180</td>\n",
              "      <td>2020-12-12</td>\n",
              "      <td>101,595.1003</td>\n",
              "      <td>82,637.0052</td>\n",
              "      <td>14,833.8982</td>\n",
              "      <td>245,277.1564</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>364</th>\n",
              "      <td>365</td>\n",
              "      <td>2021-06-15</td>\n",
              "      <td>103,243.6726</td>\n",
              "      <td>85,487.0787</td>\n",
              "      <td>14,855.1459</td>\n",
              "      <td>241,516.1694</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     day       date         mean      median          p5          p95\n",
              "0      1 2020-06-16 102,874.0184 83,968.4539 14,133.5500 253,965.0897\n",
              "179  180 2020-12-12 101,595.1003 82,637.0052 14,833.8982 245,277.1564\n",
              "364  365 2021-06-15 103,243.6726 85,487.0787 14,855.1459 241,516.1694"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "# Daily bootstrap simulation.\n",
        "bootstrap_seed = 43\n",
        "n_sims = 1_000\n",
        "horizon = 365\n",
        "pool = daily_sales[\"sales_amount\"].to_numpy(dtype=float)\n",
        "\n",
        "rng = np.random.default_rng(bootstrap_seed)\n",
        "sim_matrix = rng.choice(pool, size=(n_sims, horizon), replace=True)\n",
        "forecast_dates = pd.date_range(daily_sales[\"date\"].max() + pd.Timedelta(days=1), periods=horizon, freq=\"D\")\n",
        "\n",
        "bootstrap_daily_stats = pd.DataFrame({\n",
        "    \"day\": np.arange(1, horizon + 1),\n",
        "    \"date\": forecast_dates,\n",
        "    \"mean\": sim_matrix.mean(axis=0),\n",
        "    \"median\": np.median(sim_matrix, axis=0),\n",
        "    \"p5\": np.percentile(sim_matrix, 5, axis=0),\n",
        "    \"p95\": np.percentile(sim_matrix, 95, axis=0),\n",
        "})\n",
        "annual_totals_bootstrap = sim_matrix.sum(axis=1)\n",
        "bootstrap_annual_totals = pd.DataFrame({\n",
        "    \"simulation\": np.arange(1, n_sims + 1),\n",
        "    \"annual_total\": annual_totals_bootstrap,\n",
        "})\n",
        "bootstrap_summary = pd.DataFrame([{\n",
        "    \"method\": \"Daily Bootstrap\",\n",
        "    \"seed\": bootstrap_seed,\n",
        "    \"pool_size\": len(pool),\n",
        "    \"n_sims\": n_sims,\n",
        "    \"horizon_days\": horizon,\n",
        "    \"mean_forecast\": annual_totals_bootstrap.mean(),\n",
        "    \"p5\": np.percentile(annual_totals_bootstrap, 5),\n",
        "    \"p95\": np.percentile(annual_totals_bootstrap, 95),\n",
        "}])\n",
        "\n",
        "bootstrap_daily_stats.to_csv(CSV_DIR / \"bootstrap_daily_stats.csv\", index=False)\n",
        "bootstrap_annual_totals.to_csv(CSV_DIR / \"bootstrap_annual_totals.csv\", index=False)\n",
        "bootstrap_summary.to_csv(CSV_DIR / \"bootstrap_summary.csv\", index=False)\n",
        "bootstrap_daily_stats.loc[bootstrap_daily_stats[\"day\"].isin([1, 180, 365])].to_csv(CSV_DIR / \"bootstrap_checkpoints.csv\", index=False)\n",
        "\n",
        "print(f\"Sampling pool size: {len(pool)}\")\n",
        "print(\"Annual bootstrap summary:\")\n",
        "display(bootstrap_summary.T)\n",
        "print(\"Checkpoint daily statistics:\")\n",
        "display(bootstrap_daily_stats.loc[bootstrap_daily_stats[\"day\"].isin([1, 180, 365])])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "id": "32967635",
      "metadata": {},
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "# Bootstrap visualizations.\n",
        "fig, ax = plt.subplots(figsize=(12, 6))\n",
        "ax.plot(bootstrap_daily_stats[\"day\"], bootstrap_daily_stats[\"mean\"], color=\"#184e77\", label=\"Mean\")\n",
        "ax.plot(bootstrap_daily_stats[\"day\"], bootstrap_daily_stats[\"median\"], color=\"#d97706\", label=\"Median\")\n",
        "ax.fill_between(\n",
        "    bootstrap_daily_stats[\"day\"],\n",
        "    bootstrap_daily_stats[\"p5\"],\n",
        "    bootstrap_daily_stats[\"p95\"],\n",
        "    color=\"#76b7b2\",\n",
        "    alpha=0.28,\n",
        "    label=\"P5-P95 band\",\n",
        ")\n",
        "ax.set_title(\"Daily Bootstrap Forecast Band\")\n",
        "ax.set_xlabel(\"Forecast day\")\n",
        "ax.set_ylabel(\"Daily sales\")\n",
        "ax.legend()\n",
        "fig.tight_layout()\n",
        "fig.savefig(FIG_DIR / \"bootstrap_forecast_band.png\", dpi=160)\n",
        "plt.show()\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10, 6))\n",
        "ax.hist(annual_totals_bootstrap, bins=40, color=\"#c2410c\", alpha=0.82, edgecolor=\"white\")\n",
        "ax.axvline(annual_totals_bootstrap.mean(), color=\"#111827\", linewidth=2, label=\"Mean\")\n",
        "ax.axvline(np.percentile(annual_totals_bootstrap, 5), color=\"#b91c1c\", linestyle=\"--\", label=\"P5 / P95\")\n",
        "ax.axvline(np.percentile(annual_totals_bootstrap, 95), color=\"#b91c1c\", linestyle=\"--\")\n",
        "ax.set_title(\"Bootstrap Annual Sales Forecast\")\n",
        "ax.set_xlabel(\"Simulated annual sales\")\n",
        "ax.set_ylabel(\"Frequency\")\n",
        "ax.legend()\n",
        "fig.tight_layout()\n",
        "fig.savefig(FIG_DIR / \"bootstrap_annual_hist.png\", dpi=160)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "0fbfe3fa",
      "metadata": {},
      "source": [
        "## Model Comparison\n",
        "\n",
        "This section compares the annual distribution from Part 1 and Part 2. The contrast is central to the business interpretation: the TTM Monte Carlo model extrapolates a recent high-growth trend, while the daily bootstrap is a non-trend historical resampling baseline."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "id": "7b81d8b6",
      "metadata": {},
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Monte Carlo mean: $61,398,421.65\n",
            "Bootstrap mean: $37,103,960.45\n",
            "Mean gap: $24,294,461.20 (65.48% above bootstrap)\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>metric</th>\n",
              "      <th>Monte Carlo</th>\n",
              "      <th>Bootstrap</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Mean</td>\n",
              "      <td>61,398,421.6488</td>\n",
              "      <td>37,103,960.4518</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Std</td>\n",
              "      <td>3,788,479.5527</td>\n",
              "      <td>1,519,351.4987</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>P5</td>\n",
              "      <td>55,134,671.8075</td>\n",
              "      <td>34,633,619.5663</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>P95</td>\n",
              "      <td>67,621,311.2077</td>\n",
              "      <td>39,574,655.3738</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>CV</td>\n",
              "      <td>0.0617</td>\n",
              "      <td>0.0409</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  metric     Monte Carlo       Bootstrap\n",
              "0   Mean 61,398,421.6488 37,103,960.4518\n",
              "1    Std  3,788,479.5527  1,519,351.4987\n",
              "2     P5 55,134,671.8075 34,633,619.5663\n",
              "3    P95 67,621,311.2077 39,574,655.3738\n",
              "4     CV          0.0617          0.0409"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "# Compare Monte Carlo and Bootstrap annual forecast distributions.\n",
        "comparison = pd.DataFrame({\n",
        "    \"metric\": list(summarize(mc_simulation[\"simulated_sales\"]).keys()),\n",
        "    \"Monte Carlo\": list(summarize(mc_simulation[\"simulated_sales\"]).values()),\n",
        "    \"Bootstrap\": list(summarize(bootstrap_annual_totals[\"annual_total\"]).values()),\n",
        "})\n",
        "comparison.to_csv(CSV_DIR / \"comparison_summary.csv\", index=False)\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10, 6))\n",
        "ax.hist(mc_simulation[\"simulated_sales\"], bins=50, alpha=0.55, density=True, label=\"Monte Carlo\", color=\"#2f6f73\")\n",
        "ax.hist(bootstrap_annual_totals[\"annual_total\"], bins=40, alpha=0.55, density=True, label=\"Daily Bootstrap\", color=\"#c2410c\")\n",
        "ax.set_title(\"Annual Forecast Distribution Comparison\")\n",
        "ax.set_xlabel(\"Simulated annual sales\")\n",
        "ax.set_ylabel(\"Density\")\n",
        "ax.legend()\n",
        "fig.tight_layout()\n",
        "fig.savefig(FIG_DIR / \"comparison_histogram.png\", dpi=160)\n",
        "plt.show()\n",
        "\n",
        "mc_mean = comparison.loc[comparison[\"metric\"] == \"Mean\", \"Monte Carlo\"].iloc[0]\n",
        "boot_mean = comparison.loc[comparison[\"metric\"] == \"Mean\", \"Bootstrap\"].iloc[0]\n",
        "mean_gap = mc_mean - boot_mean\n",
        "print(f\"Monte Carlo mean: {money(mc_mean)}\")\n",
        "print(f\"Bootstrap mean: {money(boot_mean)}\")\n",
        "print(f\"Mean gap: {money(mean_gap)} ({mean_gap / boot_mean:.2%} above bootstrap)\")\n",
        "display(comparison)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "d6fd1610",
      "metadata": {},
      "source": [
        "## Category and Channel Extended Bootstrap\n",
        "\n",
        "This analysis is not a replacement for the main forecast. Its purpose is to identify whether volatility differs meaningfully across business segments. Each Category and Channel is reindexed to the complete calendar with missing segment-days filled as zero, then bootstrapped independently for 200 simulated 365-day years. The key output is CV, which supports differentiated recommendations for conservative or aggressive planning by segment."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "id": "64f17a2a",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Category bootstrap sorted by CV:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>category</th>\n",
              "      <th>history_days</th>\n",
              "      <th>total_sales</th>\n",
              "      <th>sales_share</th>\n",
              "      <th>mean</th>\n",
              "      <th>p5</th>\n",
              "      <th>p95</th>\n",
              "      <th>std</th>\n",
              "      <th>CV</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Components</td>\n",
              "      <td>1081</td>\n",
              "      <td>11,799,076.6584</td>\n",
              "      <td>0.1075</td>\n",
              "      <td>3,988,263.0798</td>\n",
              "      <td>3,543,966.5590</td>\n",
              "      <td>4,450,581.9330</td>\n",
              "      <td>286,566.3749</td>\n",
              "      <td>0.0719</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Accessories</td>\n",
              "      <td>1081</td>\n",
              "      <td>1,272,057.8878</td>\n",
              "      <td>0.0116</td>\n",
              "      <td>428,829.7971</td>\n",
              "      <td>379,118.2156</td>\n",
              "      <td>479,242.2540</td>\n",
              "      <td>30,595.9763</td>\n",
              "      <td>0.0713</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Clothing</td>\n",
              "      <td>1081</td>\n",
              "      <td>2,117,613.4491</td>\n",
              "      <td>0.0193</td>\n",
              "      <td>715,259.9036</td>\n",
              "      <td>632,937.2257</td>\n",
              "      <td>792,596.8182</td>\n",
              "      <td>48,096.2361</td>\n",
              "      <td>0.0672</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Bikes</td>\n",
              "      <td>1081</td>\n",
              "      <td>94,620,526.2077</td>\n",
              "      <td>0.8617</td>\n",
              "      <td>32,026,063.4138</td>\n",
              "      <td>29,753,967.2425</td>\n",
              "      <td>34,244,182.6734</td>\n",
              "      <td>1,326,745.6953</td>\n",
              "      <td>0.0414</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      category  history_days     total_sales  sales_share            mean  \\\n",
              "0   Components          1081 11,799,076.6584       0.1075  3,988,263.0798   \n",
              "1  Accessories          1081  1,272,057.8878       0.0116    428,829.7971   \n",
              "2     Clothing          1081  2,117,613.4491       0.0193    715,259.9036   \n",
              "3        Bikes          1081 94,620,526.2077       0.8617 32,026,063.4138   \n",
              "\n",
              "               p5             p95            std     CV  \n",
              "0  3,543,966.5590  4,450,581.9330   286,566.3749 0.0719  \n",
              "1    379,118.2156    479,242.2540    30,595.9763 0.0713  \n",
              "2    632,937.2257    792,596.8182    48,096.2361 0.0672  \n",
              "3 29,753,967.2425 34,244,182.6734 1,326,745.6953 0.0414  "
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Channel bootstrap sorted by CV:\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>channel</th>\n",
              "      <th>history_days</th>\n",
              "      <th>total_sales</th>\n",
              "      <th>sales_share</th>\n",
              "      <th>mean</th>\n",
              "      <th>p5</th>\n",
              "      <th>p95</th>\n",
              "      <th>std</th>\n",
              "      <th>CV</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Reseller</td>\n",
              "      <td>1081</td>\n",
              "      <td>80,450,596.9823</td>\n",
              "      <td>0.7326</td>\n",
              "      <td>27,252,184.3625</td>\n",
              "      <td>24,717,961.9685</td>\n",
              "      <td>29,760,846.5753</td>\n",
              "      <td>1,532,113.6239</td>\n",
              "      <td>0.0562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Internet</td>\n",
              "      <td>1081</td>\n",
              "      <td>29,358,677.2207</td>\n",
              "      <td>0.2674</td>\n",
              "      <td>9,906,231.8317</td>\n",
              "      <td>9,297,661.5482</td>\n",
              "      <td>10,422,663.1680</td>\n",
              "      <td>330,474.5473</td>\n",
              "      <td>0.0334</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    channel  history_days     total_sales  sales_share            mean  \\\n",
              "0  Reseller          1081 80,450,596.9823       0.7326 27,252,184.3625   \n",
              "1  Internet          1081 29,358,677.2207       0.2674  9,906,231.8317   \n",
              "\n",
              "               p5             p95            std     CV  \n",
              "0 24,717,961.9685 29,760,846.5753 1,532,113.6239 0.0562  \n",
              "1  9,297,661.5482 10,422,663.1680   330,474.5473 0.0334  "
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Highest-CV categories: Components, Accessories, Clothing\n",
            "Lowest-CV category: Bikes\n",
            "Highest-CV channel: Reseller\n",
            "Lowest-CV channel: Internet\n"
          ]
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "def segment_bootstrap(segment_daily: pd.DataFrame, segment_col: str, seed: int = 43, n_sims: int = 200, horizon: int = 365) -> pd.DataFrame:\n",
        "    results = []\n",
        "    overall_total = daily_sales[\"sales_amount\"].sum()\n",
        "\n",
        "    for segment_name, group in segment_daily.groupby(segment_col, sort=True):\n",
        "        group = group.sort_values(\"date\")\n",
        "        pool_segment = group[\"sales_amount\"].to_numpy(dtype=float)\n",
        "        rng = np.random.default_rng(seed)\n",
        "        sim_matrix_segment = rng.choice(pool_segment, size=(n_sims, horizon), replace=True)\n",
        "        totals = sim_matrix_segment.sum(axis=1)\n",
        "        mean_total = totals.mean()\n",
        "        std_total = totals.std(ddof=1)\n",
        "        results.append({\n",
        "            segment_col: segment_name,\n",
        "            \"history_days\": len(pool_segment),\n",
        "            \"total_sales\": pool_segment.sum(),\n",
        "            \"sales_share\": pool_segment.sum() / overall_total,\n",
        "            \"mean\": mean_total,\n",
        "            \"p5\": np.percentile(totals, 5),\n",
        "            \"p95\": np.percentile(totals, 95),\n",
        "            \"std\": std_total,\n",
        "            \"CV\": std_total / mean_total if mean_total != 0 else np.nan,\n",
        "        })\n",
        "\n",
        "    return pd.DataFrame(results).sort_values(\"CV\", ascending=False).reset_index(drop=True)\n",
        "\n",
        "\n",
        "bootstrap_by_category = segment_bootstrap(daily_category_sales, \"category\", seed=43, n_sims=200, horizon=365)\n",
        "bootstrap_by_channel = segment_bootstrap(daily_channel_sales, \"channel\", seed=43, n_sims=200, horizon=365)\n",
        "\n",
        "bootstrap_by_category.to_csv(CSV_DIR / \"bootstrap_by_category.csv\", index=False)\n",
        "bootstrap_by_channel.to_csv(CSV_DIR / \"bootstrap_by_channel.csv\", index=False)\n",
        "\n",
        "print(\"Category bootstrap sorted by CV:\")\n",
        "display(bootstrap_by_category)\n",
        "print(\"Channel bootstrap sorted by CV:\")\n",
        "display(bootstrap_by_channel)\n",
        "\n",
        "print(\"Highest-CV categories:\", \", \".join(bootstrap_by_category.head(3)[\"category\"]))\n",
        "print(\"Lowest-CV category:\", bootstrap_by_category.tail(1)[\"category\"].iloc[0])\n",
        "print(\"Highest-CV channel:\", bootstrap_by_channel.head(1)[\"channel\"].iloc[0])\n",
        "print(\"Lowest-CV channel:\", bootstrap_by_channel.tail(1)[\"channel\"].iloc[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "960d6c28",
      "metadata": {},
      "source": [
        "## Output Inventory\n",
        "\n",
        "The notebook has regenerated the main analysis outputs below. These files are used for the report and dashboard work."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "id": "38ffcfd2",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "outputs/csv/annual_sales.csv\n",
            "outputs/csv/bootstrap_annual_totals.csv\n",
            "outputs/csv/bootstrap_by_category.csv\n",
            "outputs/csv/bootstrap_by_channel.csv\n",
            "outputs/csv/bootstrap_checkpoints.csv\n",
            "outputs/csv/bootstrap_daily_stats.csv\n",
            "outputs/csv/bootstrap_summary.csv\n",
            "outputs/csv/comparison_summary.csv\n",
            "outputs/csv/daily_category_sales.csv\n",
            "outputs/csv/daily_channel_sales.csv\n",
            "outputs/csv/daily_sales.csv\n",
            "outputs/csv/data_quality_step1.csv\n",
            "outputs/csv/joined_sales_step1.csv\n",
            "outputs/csv/mc_simulation.csv\n",
            "outputs/csv/mc_summary.csv\n",
            "outputs/csv/schema_summary_step1.csv\n",
            "outputs/csv/ttm_growth.csv\n",
            "outputs/csv/ttm_sales.csv\n",
            "outputs/figures/bootstrap_annual_hist.png\n",
            "outputs/figures/bootstrap_forecast_band.png\n",
            "outputs/figures/comparison_histogram.png\n",
            "outputs/figures/mc_histogram.png\n"
          ]
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "output_files = sorted([p.relative_to(PROJECT_ROOT).as_posix() for p in CSV_DIR.glob(\"*.csv\")]) + sorted([p.relative_to(PROJECT_ROOT).as_posix() for p in FIG_DIR.glob(\"*.png\")])\n",
        "for file in output_files:\n",
        "    print(file)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "fffb90eb",
      "metadata": {},
      "source": [
        "## Key Results Audit\n",
        "\n",
        "This final cell prints the exact values needed to reconcile the notebook run against the staged results. Because the stochastic simulations now use NumPy's standard `default_rng`, Monte Carlo and Bootstrap simulation values may differ from the earlier hand-run custom-random outputs, while the deterministic TTM growth parameters should remain the same."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "id": "f39791f8",
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "AUDIT: Key Results\n",
            "TTM sample count: 12\n",
            "mu: 0.4322613770 (43.23%)\n",
            "sigma: 0.0877635375 (8.78%)\n",
            "2018 to 2019 annual YoY: 0.4056185600 (40.56%)\n",
            "\n",
            "Monte Carlo\n",
            "mean: $61,398,421.65\n",
            "p5: $55,134,671.81\n",
            "p95: $67,621,311.21\n",
            "target probability: 99.9900%\n",
            "\n",
            "Daily Bootstrap\n",
            "mean: $37,103,960.45\n",
            "p5: $34,633,619.57\n",
            "p95: $39,574,655.37\n",
            "\n",
            "Category CV ranking\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>category</th>\n",
              "      <th>mean</th>\n",
              "      <th>p5</th>\n",
              "      <th>p95</th>\n",
              "      <th>std</th>\n",
              "      <th>CV</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Components</td>\n",
              "      <td>3,988,263.0798</td>\n",
              "      <td>3,543,966.5590</td>\n",
              "      <td>4,450,581.9330</td>\n",
              "      <td>286,566.3749</td>\n",
              "      <td>0.0719</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Accessories</td>\n",
              "      <td>428,829.7971</td>\n",
              "      <td>379,118.2156</td>\n",
              "      <td>479,242.2540</td>\n",
              "      <td>30,595.9763</td>\n",
              "      <td>0.0713</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Clothing</td>\n",
              "      <td>715,259.9036</td>\n",
              "      <td>632,937.2257</td>\n",
              "      <td>792,596.8182</td>\n",
              "      <td>48,096.2361</td>\n",
              "      <td>0.0672</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Bikes</td>\n",
              "      <td>32,026,063.4138</td>\n",
              "      <td>29,753,967.2425</td>\n",
              "      <td>34,244,182.6734</td>\n",
              "      <td>1,326,745.6953</td>\n",
              "      <td>0.0414</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      category            mean              p5             p95            std  \\\n",
              "0   Components  3,988,263.0798  3,543,966.5590  4,450,581.9330   286,566.3749   \n",
              "1  Accessories    428,829.7971    379,118.2156    479,242.2540    30,595.9763   \n",
              "2     Clothing    715,259.9036    632,937.2257    792,596.8182    48,096.2361   \n",
              "3        Bikes 32,026,063.4138 29,753,967.2425 34,244,182.6734 1,326,745.6953   \n",
              "\n",
              "      CV  \n",
              "0 0.0719  \n",
              "1 0.0713  \n",
              "2 0.0672  \n",
              "3 0.0414  "
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Channel CV ranking\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>channel</th>\n",
              "      <th>mean</th>\n",
              "      <th>p5</th>\n",
              "      <th>p95</th>\n",
              "      <th>std</th>\n",
              "      <th>CV</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Reseller</td>\n",
              "      <td>27,252,184.3625</td>\n",
              "      <td>24,717,961.9685</td>\n",
              "      <td>29,760,846.5753</td>\n",
              "      <td>1,532,113.6239</td>\n",
              "      <td>0.0562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Internet</td>\n",
              "      <td>9,906,231.8317</td>\n",
              "      <td>9,297,661.5482</td>\n",
              "      <td>10,422,663.1680</td>\n",
              "      <td>330,474.5473</td>\n",
              "      <td>0.0334</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    channel            mean              p5             p95            std  \\\n",
              "0  Reseller 27,252,184.3625 24,717,961.9685 29,760,846.5753 1,532,113.6239   \n",
              "1  Internet  9,906,231.8317  9,297,661.5482 10,422,663.1680   330,474.5473   \n",
              "\n",
              "      CV  \n",
              "0 0.0562  \n",
              "1 0.0334  "
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# --- Recovery prelude ---\n",
        "from pathlib import Path\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "try:\n",
        "    import matplotlib.pyplot as plt\n",
        "    HAS_MATPLOTLIB = True\n",
        "except ModuleNotFoundError:\n",
        "    plt = None\n",
        "    HAS_MATPLOTLIB = False\n",
        "    print(\"matplotlib is not installed in this Python environment; figure cells will be skipped, but CSV/statistical outputs will still be generated.\")\n",
        "\n",
        "def _resolve_project_root():\n",
        "    start = Path.cwd()\n",
        "    for parent in [start, *start.parents]:\n",
        "        if (parent / \"outputs\" / \"csv\").exists() or (parent / \"work\" / \"AdventureWorks Sales.xlsx\").exists() or (parent / \"AdventureWorks Sales.xlsx\").exists():\n",
        "            return parent\n",
        "    return start\n",
        "\n",
        "PROJECT_ROOT = globals().get(\"PROJECT_ROOT\", _resolve_project_root())\n",
        "CSV_DIR = globals().get(\"CSV_DIR\", PROJECT_ROOT / \"outputs\" / \"csv\")\n",
        "FIG_DIR = globals().get(\"FIG_DIR\", PROJECT_ROOT / \"outputs\" / \"figures\")\n",
        "CSV_DIR.mkdir(parents=True, exist_ok=True)\n",
        "FIG_DIR.mkdir(parents=True, exist_ok=True)\n",
        "\n",
        "def money(x):\n",
        "    return \"$\" + f\"{x:,.2f}\"\n",
        "\n",
        "def summarize(values) -> dict:\n",
        "    arr = np.asarray(values, dtype=float)\n",
        "    return {\n",
        "        \"Mean\": arr.mean(),\n",
        "        \"Std\": arr.std(ddof=1),\n",
        "        \"P5\": np.percentile(arr, 5),\n",
        "        \"P95\": np.percentile(arr, 95),\n",
        "        \"CV\": arr.std(ddof=1) / arr.mean() if arr.mean() != 0 else np.nan,\n",
        "    }\n",
        "\n",
        "def _load_core_csv_if_needed():\n",
        "    global daily_sales, annual_sales, daily_category_sales, daily_channel_sales\n",
        "    if \"daily_sales\" not in globals() and (CSV_DIR / \"daily_sales.csv\").exists():\n",
        "        daily_sales = pd.read_csv(CSV_DIR / \"daily_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"annual_sales\" not in globals() and (CSV_DIR / \"annual_sales.csv\").exists():\n",
        "        annual_sales = pd.read_csv(CSV_DIR / \"annual_sales.csv\")\n",
        "    if \"daily_category_sales\" not in globals() and (CSV_DIR / \"daily_category_sales.csv\").exists():\n",
        "        daily_category_sales = pd.read_csv(CSV_DIR / \"daily_category_sales.csv\", parse_dates=[\"date\"])\n",
        "    if \"daily_channel_sales\" not in globals() and (CSV_DIR / \"daily_channel_sales.csv\").exists():\n",
        "        daily_channel_sales = pd.read_csv(CSV_DIR / \"daily_channel_sales.csv\", parse_dates=[\"date\"])\n",
        "\n",
        "_load_core_csv_if_needed()\n",
        "# --- End recovery prelude ---\n",
        "if \"ttm_growth\" not in globals() and (CSV_DIR / \"ttm_growth.csv\").exists():\n",
        "    ttm_growth = pd.read_csv(CSV_DIR / \"ttm_growth.csv\", parse_dates=[\"month_end\"])\n",
        "    mu = ttm_growth[\"ttm_growth\"].mean()\n",
        "    sigma = ttm_growth[\"ttm_growth\"].std(ddof=1)\n",
        "if \"annual_yoy_2018_to_2019\" not in globals() and \"annual_sales\" in globals():\n",
        "    annual_yoy_2018_to_2019 = (annual_sales.loc[annual_sales[\"year\"] == 2019, \"sales_amount\"].iloc[0] - annual_sales.loc[annual_sales[\"year\"] == 2018, \"sales_amount\"].iloc[0]) / annual_sales.loc[annual_sales[\"year\"] == 2018, \"sales_amount\"].iloc[0]\n",
        "if \"mc_summary\" not in globals() and (CSV_DIR / \"mc_summary.csv\").exists():\n",
        "    mc_summary = pd.read_csv(CSV_DIR / \"mc_summary.csv\")\n",
        "if \"mc_simulation\" not in globals() and (CSV_DIR / \"mc_simulation.csv\").exists():\n",
        "    mc_simulation = pd.read_csv(CSV_DIR / \"mc_simulation.csv\")\n",
        "if \"bootstrap_summary\" not in globals() and (CSV_DIR / \"bootstrap_summary.csv\").exists():\n",
        "    bootstrap_summary = pd.read_csv(CSV_DIR / \"bootstrap_summary.csv\")\n",
        "if \"bootstrap_annual_totals\" not in globals() and (CSV_DIR / \"bootstrap_annual_totals.csv\").exists():\n",
        "    bootstrap_annual_totals = pd.read_csv(CSV_DIR / \"bootstrap_annual_totals.csv\")\n",
        "if \"comparison\" not in globals() and (CSV_DIR / \"comparison_summary.csv\").exists():\n",
        "    comparison = pd.read_csv(CSV_DIR / \"comparison_summary.csv\")\n",
        "if \"bootstrap_by_category\" not in globals() and (CSV_DIR / \"bootstrap_by_category.csv\").exists():\n",
        "    bootstrap_by_category = pd.read_csv(CSV_DIR / \"bootstrap_by_category.csv\")\n",
        "if \"bootstrap_by_channel\" not in globals() and (CSV_DIR / \"bootstrap_by_channel.csv\").exists():\n",
        "    bootstrap_by_channel = pd.read_csv(CSV_DIR / \"bootstrap_by_channel.csv\")\n",
        "\n",
        "print(\"AUDIT: Key Results\")\n",
        "print(f\"TTM sample count: {len(ttm_growth)}\")\n",
        "print(f\"mu: {mu:.10f} ({mu:.2%})\")\n",
        "print(f\"sigma: {sigma:.10f} ({sigma:.2%})\")\n",
        "print(f\"2018 to 2019 annual YoY: {annual_yoy_2018_to_2019:.10f} ({annual_yoy_2018_to_2019:.2%})\")\n",
        "\n",
        "mc_from_summary = float(mc_summary.iloc[0][\"mean_forecast\"])\n",
        "mc_from_sim = float(mc_simulation[\"simulated_sales\"].mean())\n",
        "mc_from_comparison = float(comparison.loc[comparison[\"metric\"] == \"Mean\", \"Monte Carlo\"].iloc[0]) if \"comparison\" in globals() else np.nan\n",
        "print(\"\\nMonte Carlo consistency check\")\n",
        "print(f\"mc_summary mean_forecast: {money(mc_from_summary)}\")\n",
        "print(f\"mc_simulation recomputed mean: {money(mc_from_sim)}\")\n",
        "print(f\"comparison Mean: {money(mc_from_comparison)}\")\n",
        "print(f\"difference summary - simulation: {mc_from_summary - mc_from_sim:,.6f}\")\n",
        "\n",
        "mc_stats = summarize(mc_simulation[\"simulated_sales\"])\n",
        "print(\"\\nMonte Carlo final values from mc_simulation\")\n",
        "print(f\"mean: {money(mc_stats['Mean'])}\")\n",
        "print(f\"std: {money(mc_stats['Std'])}\")\n",
        "print(f\"p5: {money(mc_stats['P5'])}\")\n",
        "print(f\"p95: {money(mc_stats['P95'])}\")\n",
        "print(f\"CV: {mc_stats['CV']:.4%}\")\n",
        "print(f\"target probability: {float(mc_summary.iloc[0]['target_probability']):.4%}\")\n",
        "\n",
        "boot_stats = summarize(bootstrap_annual_totals[\"annual_total\"])\n",
        "print(\"\\nDaily Bootstrap\")\n",
        "print(f\"mean: {money(boot_stats['Mean'])}\")\n",
        "print(f\"std: {money(boot_stats['Std'])}\")\n",
        "print(f\"p5: {money(boot_stats['P5'])}\")\n",
        "print(f\"p95: {money(boot_stats['P95'])}\")\n",
        "print(f\"CV: {boot_stats['CV']:.4%}\")\n",
        "\n",
        "print(\"\\nCategory CV ranking\")\n",
        "display(bootstrap_by_category[[\"category\", \"mean\", \"p5\", \"p95\", \"std\", \"CV\"]])\n",
        "\n",
        "print(\"Channel CV ranking\")\n",
        "display(bootstrap_by_channel[[\"channel\", \"mean\", \"p5\", \"p95\", \"std\", \"CV\"]])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "e4daa3d2-a059-4d0f-8f2d-3c51e109b150",
      "metadata": {},
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "4b14b091-1a3c-4085-a6dd-4b8ddb06e5b0",
      "metadata": {},
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "c3ace22d-7f47-46ec-a73a-5a9800cf909d",
      "metadata": {},
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.14.3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}