{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c462b967",
   "metadata": {},
   "source": [
    "# One-Factor Plots\n",
    "\n",
    "One-factor EDA assumes the model **y = f(x) + e**.  `x` is typically a\n",
    "categorical factor (batch, machine, operator, treatment group) and the\n",
    "plots answer: *do the factor levels produce different location or spread?*\n",
    "\n",
    "Pass `x` to `EDAData` as a 1-D array of labels or numbers, matching `len(y)`.\n",
    "\n",
    "Reference: [NIST Handbook Chapter 1.3.3](https://www.itl.nist.gov/div898/handbook/eda/section3/eda33.htm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5221754",
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from drippy import EDAData\n",
    "from drippy import (\n",
    "    scatter_plot,\n",
    "    box_plot,\n",
    "    bihistogram,\n",
    "    qq_plot,\n",
    "    mean_plot,\n",
    "    sd_plot,\n",
    ")\n",
    "\n",
    "rng = np.random.default_rng(42)\n",
    "\n",
    "# Five production batches — 20 observations each\n",
    "batches = [\"B1\", \"B2\", \"B3\", \"B4\", \"B5\"]\n",
    "x_cat = np.repeat(batches, 20)\n",
    "locs = [5.0, 5.5, 4.8, 5.2, 5.1]\n",
    "y_cat = np.concatenate(\n",
    "    [rng.normal(loc=l, scale=1.0, size=20) for l in locs]\n",
    ")\n",
    "data_cat = EDAData(y=y_cat, x=x_cat)\n",
    "\n",
    "# Two-group comparison — 30 observations each\n",
    "x_two = np.repeat([\"Method_A\", \"Method_B\"], 30)\n",
    "y_two = np.concatenate(\n",
    "    [rng.normal(10.0, 1.0, 30), rng.normal(10.5, 1.2, 30)]\n",
    ")\n",
    "data_two = EDAData(y=y_two, x=x_two)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8aa8929f",
   "metadata": {},
   "source": [
    "## Scatter Plot  (NIST 1.3.3.26)\n",
    "\n",
    "Raw scatter of `y` vs `x`.  Works with both continuous and categorical `x`.\n",
    "The simplest way to spot location differences, outliers, or non-linearity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "33dbfe59",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_cont = np.linspace(1, 5, 100)\n",
    "y_cont = 2.0 * x_cont + rng.normal(scale=0.5, size=100)\n",
    "data_cont = EDAData(y=y_cont, x=x_cont)\n",
    "\n",
    "fig, ax = scatter_plot(data_cont)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0573ac5",
   "metadata": {},
   "source": [
    "## Box Plot  (NIST 1.3.3.7)\n",
    "\n",
    "Displays median, interquartile range, and outliers for each factor level.\n",
    "Ideal for comparing location and spread across many groups simultaneously."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d32ebf90",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, ax = box_plot(data_cat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b80fa36",
   "metadata": {},
   "source": [
    "## Bihistogram  (NIST 1.3.3.2)\n",
    "\n",
    "Side-by-side histograms for exactly two factor levels.  The mirrored\n",
    "layout makes it easy to compare distribution shape, location, and spread\n",
    "between, for example, two measurement methods or two labs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be275bfe",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, axes = bihistogram(data_two)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "093f16de",
   "metadata": {},
   "source": [
    "## Q-Q Plot  (NIST 1.3.3.23)\n",
    "\n",
    "Quantile-quantile plot comparing two groups.  Points on the diagonal\n",
    "indicate identical distributions; vertical shifts indicate location\n",
    "differences; changes in slope indicate scale differences."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a7ccfc69",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, ax = qq_plot(data_two)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b1ded59",
   "metadata": {},
   "source": [
    "## Mean Plot  (NIST 1.3.3.17)\n",
    "\n",
    "Group means connected by a line, with a grand-mean reference.\n",
    "A flat line indicates no factor effect; visible trends or steps\n",
    "suggest systematic differences between levels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5e27dbd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, ax = mean_plot(data_cat)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5bf35564",
   "metadata": {},
   "source": [
    "## SD Plot  (NIST 1.3.3.28)\n",
    "\n",
    "Group standard deviations connected by a line, with a pooled-SD\n",
    "reference.  Detects heteroscedasticity — factor levels with unusually\n",
    "high or low variability."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "30068a0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, ax = sd_plot(data_cat)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}