doc/examples/py_double_ml_multiway_cluster.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "mNcwBqn11_ZX"
   },
   "source": [
    "# Python: Cluster Robust Double Machine Learning\n",
    "\n",
    "\n",
    "## Motivation \n",
    "\n",
    "In many empirical applications, errors exhibit a clustered structure such that the usual i.i.d. assumption does not hold anymore. In order to perform valid statistical inference, researchers have to account for clustering. In this notebook, we will shortly emphasize the consequences of clustered data on inference based on the double machine learning (DML) approach as has been considered in [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815). We will demonstrate how users of the [DoubleML](https://docs.doubleml.org/stable/index.html) package can account for one- and two-way clustering in their analysis.\n",
    "\n",
    "Clustered errors in terms of one or multiple dimensions might arise in many empirical applications. For example, in a cross-sectional study, errors might be correlated (i) within regions (one-way clustering) or (ii) within regions and industries at the same time (two-way clustering). Another example for two-way clustering, discussed in [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815), refers to market share data with market shares being subject to shocks on the market and product level at the same time. We refer to [Cameron et al. (2011)](https://doi.org/10.1198/jbes.2010.07136) for an introduction to multiway clustering and a illustrative list of empirical examples.\n",
    "\n",
    "## Clustering and double machine learning\n",
    "\n",
    "Clustering creates a challenge to the double machine learning (DML) approach in terms of \n",
    "\n",
    "1. a necessary adjustment of the formulae used for estimation of the variance covariance matrix, standard errors, p-values etc., and,\n",
    "2. an adjusted resampling scheme for the cross-fitting algorithm.\n",
    "\n",
    "The first point equally applies to classical statistical models, for example a linear regression model (see, for example [Cameron et al. 2011](https://doi.org/10.1198/jbes.2010.07136)). The second point arises because the clustering implies a correlation of errors from train and test samples if the standard cross-fitting procedure suggested in [Chernozhukov et al. (2018)](https://doi.org/10.1111/ectj.12097) was employed. The DML approach builds on independent sample splits into partitions that are used for training of the machine learning (ML) model learners and generation of predictions that are eventually used for solving the score function. For a motivation of the necessity of sample splitting, we refer to the illustration example in the [user guide]( \n",
    "https://docs.doubleml.org/stable/guide/basics.html#sample-splitting-to-remove-bias-induced-by-overfitting) as well as to the explanation in [Chernozhukov et al. (2018)](https://doi.org/10.1111/ectj.12097) .\n",
    "\n",
    "In order to achieve independent data splits in a setting with one-way or multi-way clustering, [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815) develop an updated  $K$-fold sample splitting procedure that ensures independent sample splits: The data set is split into disjoint partitions in terms of all clustering dimensions. For example, in a situation with two-way clustering, the data is split into $K^2$ folds. The machine learning models are then trained on a specific fold and used for generation of predictions in hold-out samples. Thereby, the sample splitting procedure ensures that the hold-out samples do not contain observations of the same clusters as used for training.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "r1nf2-YD1_Z2"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.colors import ListedColormap\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.model_selection import KFold, RepeatedKFold\n",
    "from sklearn.base import clone\n",
    "\n",
    "from sklearn.linear_model import LassoCV\n",
    "\n",
    "from doubleml import DoubleMLClusterData, DoubleMLData, DoubleMLPLIV\n",
    "\n",
    "from doubleml.datasets import make_pliv_multiway_cluster_CKMS2021"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "1lptyTy61_Z_"
   },
   "source": [
    "## A Motivating Example: Two-Way Cluster Robust DML\n",
    "\n",
    "In a first part, we show how the two-way cluster robust double machine learning (DML) ([Chiang et al. 2021](https://doi.org/10.1080/07350015.2021.1895815)) can be implemented with the [DoubleML](https://docs.doubleml.org/stable/index.html) package.\n",
    "[Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815) consider double-indexed data\n",
    "\n",
    "\\begin{equation}\n",
    "\\lbrace W_{ij}: i \\in \\lbrace 1, \\ldots, N \\rbrace, j \\in \\lbrace 1, \\ldots, M \\rbrace \\rbrace\n",
    "\\end{equation}\n",
    "\n",
    "and the partially linear IV regression model (PLIV)\n",
    "\n",
    "$$\\begin{aligned}\n",
    "Y_{ij} = D_{ij} \\theta_0 +  g_0(X_{ij}) + \\epsilon_{ij}, & &\\mathbb{E}(\\epsilon_{ij} | X_{ij}, Z_{ij}) = 0, \\\\\n",
    "Z_{ij} = m_0(X_{ij}) + v_{ij}, & &\\mathbb{E}(v_{ij} | X_{ij}) = 0.\n",
    "\\end{aligned}$$"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "9L-nrVMH1_aD"
   },
   "source": [
    "### Simulate two-way cluster data\n",
    "\n",
    "We use the PLIV data generating process described in Section 4.1 of [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815).\n",
    "The DGP is defined as\n",
    "$$\\begin{aligned}\n",
    "Z_{ij} &= X_{ij}' \\xi_0 + V_{ij}, \\\\\n",
    "D_{ij} &= Z_{ij}' \\pi_{10} + X_{ij}' \\pi_{20} + v_{ij}, \\\\\n",
    "Y_{ij} &= D_{ij} \\theta + X_{ij}' \\zeta_0 + \\varepsilon_{ij},\n",
    "\\end{aligned}$$\n",
    "with\n",
    "$$\\begin{aligned}\n",
    "X_{ij} &= (1 - \\omega_1^X - \\omega_2^X) \\alpha_{ij}^X\n",
    "+ \\omega_1^X \\alpha_{i}^X + \\omega_2^X \\alpha_{j}^X, \\\\\n",
    "\\varepsilon_{ij} &= (1 - \\omega_1^\\varepsilon - \\omega_2^\\varepsilon) \\alpha_{ij}^\\varepsilon\n",
    "+ \\omega_1^\\varepsilon \\alpha_{i}^\\varepsilon + \\omega_2^\\varepsilon \\alpha_{j}^\\varepsilon, \\\\\n",
    "v_{ij} &= (1 - \\omega_1^v - \\omega_2^v) \\alpha_{ij}^v\n",
    "+ \\omega_1^v \\alpha_{i}^v + \\omega_2^v \\alpha_{j}^v, \\\\\n",
    "V_{ij} &= (1 - \\omega_1^V - \\omega_2^V) \\alpha_{ij}^V\n",
    "+ \\omega_1^V \\alpha_{i}^V + \\omega_2^V \\alpha_{j}^V,\n",
    "\\end{aligned}$$\n",
    "and $\\alpha_{ij}^X, \\alpha_{i}^X, \\alpha_{j}^X \\sim \\mathcal{N}(0, \\Sigma)$\n",
    "where  $\\Sigma$ is a $p_x \\times p_x$ matrix with entries\n",
    "$\\Sigma_{kj} = s_X^{|j-k|}$.\n",
    "Further\n",
    "$$\\begin{aligned}\n",
    "\\left(\\begin{matrix} \\alpha_{ij}^\\varepsilon \\\\ \\alpha_{ij}^v \\end{matrix}\\right),\n",
    "\\left(\\begin{matrix} \\alpha_{i}^\\varepsilon \\\\ \\alpha_{i}^v \\end{matrix}\\right),\n",
    "\\left(\\begin{matrix} \\alpha_{j}^\\varepsilon \\\\ \\alpha_{j}^v \\end{matrix}\\right)\n",
    "\\sim \\mathcal{N}\\left(0, \\left(\\begin{matrix} 1 & s_{\\varepsilon v} \\\\\n",
    "s_{\\varepsilon v} & 1 \\end{matrix} \\right) \\right)\n",
    "\\end{aligned}$$\n",
    "and $\\alpha_{ij}^V, \\alpha_{i}^V, \\alpha_{j}^V \\sim \\mathcal{N}(0, 1)$.\n",
    "\n",
    "Data from this DGP can be generated with the [make_pliv_multiway_cluster_CKMS2021()](https://docs.doubleml.org/stable/api/generated/doubleml.datasets.make_pliv_multiway_cluster_CKMS2021.html#doubleml.datasets.make_pliv_multiway_cluster_CKMS2021) function from [DoubleML](https://docs.doubleml.org/stable/index.html).\n",
    "Analogously to [Chiang et al. (2021, Section 5)](https://doi.org/10.1080/07350015.2021.1895815)\n",
    "we use the following parameter setting:\n",
    "$\\theta=1.0$, $N=M=25$, $p_x=100$, $\\pi_{10}=1.0$, $\\omega_X = \\omega_{\\varepsilon} = \\omega_V = \\omega_v = (0.25, 0.25)$, $s_X = s_{\\varepsilon v} = 0.25$ and the $j$-th entries of the $p_x$-vectors $\\zeta_0 = \\pi_{20} = \\xi_0$ are $(\\zeta_{0})_j = 0.5^j$.\n",
    "This are also the default values of [make_pliv_multiway_cluster_CKMS2021()](https://docs.doubleml.org/stable/api/generated/doubleml.datasets.make_pliv_multiway_cluster_CKMS2021.html#doubleml.datasets.make_pliv_multiway_cluster_CKMS2021)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "9hepXKz91_aM"
   },
   "outputs": [],
   "source": [
    "# Set the simulation parameters\n",
    "N = 25  # number of observations (first dimension)\n",
    "M = 25  # number of observations (second dimension)\n",
    "dim_X = 100  # dimension of X\n",
    "np.random.seed(3141) # set seed\n",
    "\n",
    "obj_dml_data = make_pliv_multiway_cluster_CKMS2021(N, M, dim_X)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "rlePOlu21_aP"
   },
   "source": [
    "### Data-Backend for Cluster Data\n",
    "The implementation of cluster robust double machine learning is based on a special data-backend called [DoubleMLClusterData](https://docs.doubleml.org/stable/api/generated/doubleml.DoubleMLClusterData.html#doubleml.DoubleMLClusterData). As compared to the standard data-backend [DoubleMLData](https://docs.doubleml.org/dev/api/generated/doubleml.DoubleMLData.html), users can specify the clustering variables during instantiation of a  [DoubleMLClusterData](https://docs.doubleml.org/stable/api/generated/doubleml.DoubleMLClusterData.html#doubleml.DoubleMLClusterData) object. The estimation framework will subsequently account for the provided clustering options."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "N9DBwEFT1_aS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== DoubleMLClusterData Object ==================\n",
      "\n",
      "------------------ Data summary      ------------------\n",
      "Outcome variable: Y\n",
      "Treatment variable(s): ['D']\n",
      "Cluster variable(s): ['cluster_var_i', 'cluster_var_j']\n",
      "Covariates: ['X1', 'X2', 'X3', 'X4', 'X5', 'X6', 'X7', 'X8', 'X9', 'X10', 'X11', 'X12', 'X13', 'X14', 'X15', 'X16', 'X17', 'X18', 'X19', 'X20', 'X21', 'X22', 'X23', 'X24', 'X25', 'X26', 'X27', 'X28', 'X29', 'X30', 'X31', 'X32', 'X33', 'X34', 'X35', 'X36', 'X37', 'X38', 'X39', 'X40', 'X41', 'X42', 'X43', 'X44', 'X45', 'X46', 'X47', 'X48', 'X49', 'X50', 'X51', 'X52', 'X53', 'X54', 'X55', 'X56', 'X57', 'X58', 'X59', 'X60', 'X61', 'X62', 'X63', 'X64', 'X65', 'X66', 'X67', 'X68', 'X69', 'X70', 'X71', 'X72', 'X73', 'X74', 'X75', 'X76', 'X77', 'X78', 'X79', 'X80', 'X81', 'X82', 'X83', 'X84', 'X85', 'X86', 'X87', 'X88', 'X89', 'X90', 'X91', 'X92', 'X93', 'X94', 'X95', 'X96', 'X97', 'X98', 'X99', 'X100']\n",
      "Instrument variable(s): ['Z']\n",
      "No. Observations: 625\n",
      "\n",
      "------------------ DataFrame info    ------------------\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 625 entries, 0 to 624\n",
      "Columns: 105 entries, cluster_var_i to Z\n",
      "dtypes: float64(103), int64(2)\n",
      "memory usage: 512.8 KB\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# The simulated data is of type DoubleMLClusterData\n",
    "print(obj_dml_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "yMU5Y0n81_aW"
   },
   "outputs": [
    {
     "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>cluster_var_i</th>\n",
       "      <th>cluster_var_j</th>\n",
       "      <th>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>X3</th>\n",
       "      <th>X4</th>\n",
       "      <th>X5</th>\n",
       "      <th>X6</th>\n",
       "      <th>X7</th>\n",
       "      <th>X8</th>\n",
       "      <th>...</th>\n",
       "      <th>X94</th>\n",
       "      <th>X95</th>\n",
       "      <th>X96</th>\n",
       "      <th>X97</th>\n",
       "      <th>X98</th>\n",
       "      <th>X99</th>\n",
       "      <th>X100</th>\n",
       "      <th>Y</th>\n",
       "      <th>D</th>\n",
       "      <th>Z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-0.261903</td>\n",
       "      <td>-0.195564</td>\n",
       "      <td>-0.118952</td>\n",
       "      <td>0.508459</td>\n",
       "      <td>0.226598</td>\n",
       "      <td>-0.544555</td>\n",
       "      <td>0.183888</td>\n",
       "      <td>1.694919</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.205938</td>\n",
       "      <td>-0.996934</td>\n",
       "      <td>-1.136836</td>\n",
       "      <td>0.010269</td>\n",
       "      <td>-0.396985</td>\n",
       "      <td>-0.161141</td>\n",
       "      <td>-0.614188</td>\n",
       "      <td>0.256567</td>\n",
       "      <td>-0.113780</td>\n",
       "      <td>0.104787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-1.082973</td>\n",
       "      <td>-0.303835</td>\n",
       "      <td>0.760778</td>\n",
       "      <td>-0.542671</td>\n",
       "      <td>-0.601598</td>\n",
       "      <td>-0.201768</td>\n",
       "      <td>0.234910</td>\n",
       "      <td>0.212844</td>\n",
       "      <td>...</td>\n",
       "      <td>0.427573</td>\n",
       "      <td>0.303324</td>\n",
       "      <td>0.247826</td>\n",
       "      <td>0.109273</td>\n",
       "      <td>-0.410795</td>\n",
       "      <td>-0.128408</td>\n",
       "      <td>0.633433</td>\n",
       "      <td>-1.816318</td>\n",
       "      <td>-1.002983</td>\n",
       "      <td>-0.942661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>-0.110359</td>\n",
       "      <td>-0.679539</td>\n",
       "      <td>0.491245</td>\n",
       "      <td>-0.309772</td>\n",
       "      <td>-0.552727</td>\n",
       "      <td>0.036729</td>\n",
       "      <td>-0.673302</td>\n",
       "      <td>-1.024604</td>\n",
       "      <td>...</td>\n",
       "      <td>0.094381</td>\n",
       "      <td>-0.922996</td>\n",
       "      <td>-2.054068</td>\n",
       "      <td>-0.477474</td>\n",
       "      <td>-0.543380</td>\n",
       "      <td>-0.720664</td>\n",
       "      <td>-0.332996</td>\n",
       "      <td>-0.851366</td>\n",
       "      <td>-0.302648</td>\n",
       "      <td>0.628069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>-0.010940</td>\n",
       "      <td>1.003427</td>\n",
       "      <td>0.412653</td>\n",
       "      <td>0.784238</td>\n",
       "      <td>0.014637</td>\n",
       "      <td>0.611269</td>\n",
       "      <td>0.323679</td>\n",
       "      <td>-0.557999</td>\n",
       "      <td>...</td>\n",
       "      <td>0.698694</td>\n",
       "      <td>1.342992</td>\n",
       "      <td>-1.136089</td>\n",
       "      <td>-0.632058</td>\n",
       "      <td>-0.509958</td>\n",
       "      <td>-0.456370</td>\n",
       "      <td>-0.557595</td>\n",
       "      <td>1.081488</td>\n",
       "      <td>0.438960</td>\n",
       "      <td>0.054162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>-0.073207</td>\n",
       "      <td>-0.370736</td>\n",
       "      <td>-0.005857</td>\n",
       "      <td>-0.833907</td>\n",
       "      <td>-0.096337</td>\n",
       "      <td>-0.714240</td>\n",
       "      <td>0.094026</td>\n",
       "      <td>0.435401</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.149714</td>\n",
       "      <td>-0.164864</td>\n",
       "      <td>-0.756805</td>\n",
       "      <td>0.518175</td>\n",
       "      <td>0.510385</td>\n",
       "      <td>-0.681176</td>\n",
       "      <td>-1.020271</td>\n",
       "      <td>-1.708190</td>\n",
       "      <td>-1.805007</td>\n",
       "      <td>-0.850321</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 105 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   cluster_var_i  cluster_var_j        X1        X2        X3        X4  \\\n",
       "0              0              0 -0.261903 -0.195564 -0.118952  0.508459   \n",
       "1              0              1 -1.082973 -0.303835  0.760778 -0.542671   \n",
       "2              0              2 -0.110359 -0.679539  0.491245 -0.309772   \n",
       "3              0              3 -0.010940  1.003427  0.412653  0.784238   \n",
       "4              0              4 -0.073207 -0.370736 -0.005857 -0.833907   \n",
       "\n",
       "         X5        X6        X7        X8  ...       X94       X95       X96  \\\n",
       "0  0.226598 -0.544555  0.183888  1.694919  ... -0.205938 -0.996934 -1.136836   \n",
       "1 -0.601598 -0.201768  0.234910  0.212844  ...  0.427573  0.303324  0.247826   \n",
       "2 -0.552727  0.036729 -0.673302 -1.024604  ...  0.094381 -0.922996 -2.054068   \n",
       "3  0.014637  0.611269  0.323679 -0.557999  ...  0.698694  1.342992 -1.136089   \n",
       "4 -0.096337 -0.714240  0.094026  0.435401  ... -0.149714 -0.164864 -0.756805   \n",
       "\n",
       "        X97       X98       X99      X100         Y         D         Z  \n",
       "0  0.010269 -0.396985 -0.161141 -0.614188  0.256567 -0.113780  0.104787  \n",
       "1  0.109273 -0.410795 -0.128408  0.633433 -1.816318 -1.002983 -0.942661  \n",
       "2 -0.477474 -0.543380 -0.720664 -0.332996 -0.851366 -0.302648  0.628069  \n",
       "3 -0.632058 -0.509958 -0.456370 -0.557595  1.081488  0.438960  0.054162  \n",
       "4  0.518175  0.510385 -0.681176 -1.020271 -1.708190 -1.805007 -0.850321  \n",
       "\n",
       "[5 rows x 105 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The cluster variables are part of the DataFrame\n",
    "obj_dml_data.data.head()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "hPXEpA3J1_aa"
   },
   "source": [
    "### Initialize the objects of class `DoubleMLPLIV`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "DvjBpafF1_ae"
   },
   "outputs": [],
   "source": [
    "# Set machine learning methods for l, m & r\n",
    "learner = LassoCV()\n",
    "ml_l = clone(learner)\n",
    "ml_m = clone(learner)\n",
    "ml_r = clone(learner)\n",
    "\n",
    "# initialize the DoubleMLPLIV object\n",
    "dml_pliv_obj = DoubleMLPLIV(obj_dml_data,\n",
    "                            ml_l, ml_m, ml_r,\n",
    "                            n_folds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "rh3-MK2Q1_ah"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== DoubleMLPLIV Object ==================\n",
      "\n",
      "------------------ Data summary      ------------------\n",
      "Outcome variable: Y\n",
      "Treatment variable(s): ['D']\n",
      "Cluster variable(s): ['cluster_var_i', 'cluster_var_j']\n",
      "Covariates: ['X1', 'X2', 'X3', 'X4', 'X5', 'X6', 'X7', 'X8', 'X9', 'X10', 'X11', 'X12', 'X13', 'X14', 'X15', 'X16', 'X17', 'X18', 'X19', 'X20', 'X21', 'X22', 'X23', 'X24', 'X25', 'X26', 'X27', 'X28', 'X29', 'X30', 'X31', 'X32', 'X33', 'X34', 'X35', 'X36', 'X37', 'X38', 'X39', 'X40', 'X41', 'X42', 'X43', 'X44', 'X45', 'X46', 'X47', 'X48', 'X49', 'X50', 'X51', 'X52', 'X53', 'X54', 'X55', 'X56', 'X57', 'X58', 'X59', 'X60', 'X61', 'X62', 'X63', 'X64', 'X65', 'X66', 'X67', 'X68', 'X69', 'X70', 'X71', 'X72', 'X73', 'X74', 'X75', 'X76', 'X77', 'X78', 'X79', 'X80', 'X81', 'X82', 'X83', 'X84', 'X85', 'X86', 'X87', 'X88', 'X89', 'X90', 'X91', 'X92', 'X93', 'X94', 'X95', 'X96', 'X97', 'X98', 'X99', 'X100']\n",
      "Instrument variable(s): ['Z']\n",
      "No. Observations: 625\n",
      "\n",
      "------------------ Score & algorithm ------------------\n",
      "Score function: partialling out\n",
      "DML algorithm: dml2\n",
      "\n",
      "------------------ Machine learner   ------------------\n",
      "Learner ml_l: LassoCV()\n",
      "Learner ml_m: LassoCV()\n",
      "Learner ml_r: LassoCV()\n",
      "\n",
      "------------------ Resampling        ------------------\n",
      "No. folds per cluster: 3\n",
      "No. folds: 9\n",
      "No. repeated sample splits: 1\n",
      "Apply cross-fitting: True\n",
      "\n",
      "------------------ Fit summary       ------------------\n",
      "Empty DataFrame\n",
      "Columns: [coef, std err, t, P>|t|]\n",
      "Index: []\n"
     ]
    }
   ],
   "source": [
    "print(dml_pliv_obj)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "nbsphinx": "hidden"
   },
   "outputs": [],
   "source": [
    "#discrete color scheme\n",
    "x = sns.color_palette(\"RdBu_r\", 7)\n",
    "cMap = ListedColormap([x[0], x[3], x[6]])\n",
    "plt.rcParams['figure.figsize'] = 15, 12\n",
    "sns.set(font_scale=1.3)\n",
    "\n",
    "def plt_smpls(smpls, n_folds_per_cluster):\n",
    "    df = pd.DataFrame(np.zeros([N*M, n_folds_per_cluster*n_folds_per_cluster]))\n",
    "    for i_split, this_split_ind in enumerate(smpls):\n",
    "        df.loc[this_split_ind[0], i_split] = -1.\n",
    "        df.loc[this_split_ind[1], i_split] = 1.\n",
    "\n",
    "    ax = sns.heatmap(df, cmap=cMap);\n",
    "    ax.invert_yaxis();\n",
    "    ax.set_ylim([0, N*M]);\n",
    "    ax.set_xlabel('Fold')\n",
    "    ax.set_ylabel('Observation')\n",
    "    colorbar = ax.collections[0].colorbar\n",
    "    colorbar.set_ticks([-0.667, 0, 0.667])\n",
    "    colorbar.set_ticklabels(['Nuisance', '', 'Score'])\n",
    "\n",
    "def plt_smpls_cluster(smpls_cluster, n_folds_per_cluster):\n",
    "    for i_split in range(len(smpls_cluster)):\n",
    "        plt.subplot(n_folds_per_cluster, n_folds_per_cluster, i_split + 1)\n",
    "        df = pd.DataFrame(np.zeros([N*M, 1]),\n",
    "                      index = pd.MultiIndex.from_product([range(N), range(M)]),\n",
    "                      columns=['value'])\n",
    "\n",
    "        df.loc[pd.MultiIndex.from_product(smpls_cluster[i_split][0]), :] = -1.\n",
    "        df.loc[pd.MultiIndex.from_product(smpls_cluster[i_split][1]), :] = 1.\n",
    "\n",
    "        df_wide = df.reset_index().pivot(index=\"level_0\", columns=\"level_1\", values=\"value\")\n",
    "        df_wide.index.name=''\n",
    "        df_wide.columns.name=''\n",
    "\n",
    "        ax = sns.heatmap(df_wide, cmap=cMap);\n",
    "        ax.invert_yaxis();\n",
    "        ax.set_ylim([0, M]);\n",
    "        colorbar = ax.collections[0].colorbar\n",
    "        colorbar.set_ticks([-0.667, 0, 0.667])\n",
    "\n",
    "        l = i_split % n_folds_per_cluster + 1\n",
    "        k = np.floor_divide(i_split, n_folds_per_cluster) + 1\n",
    "        title = f'Nuisance: $I_{{{k}}}^C \\\\times J_{{{l}}}^C$; Score: $I_{{{k}}} \\\\times J_{{{l}}}$'\n",
    "        ax.set_title(title)\n",
    "        if l == n_folds_per_cluster:\n",
    "            colorbar.set_ticklabels(['Nuisance', '', 'Score'])\n",
    "        else:\n",
    "            colorbar.set_ticklabels(['', '', ''])\n",
    "        if l == 1:\n",
    "            ax.set_ylabel('First Cluster Variable $k$')\n",
    "        if k == 3:\n",
    "            ax.set_xlabel('Second Cluster Variable $\\ell$')\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "dbxf265L1_ak"
   },
   "source": [
    "### Cluster Robust Cross Fitting\n",
    "A key element of cluster robust DML ([Chiang et al. 2021](https://doi.org/10.1080/07350015.2021.1895815)) is a special sample splitting used for the cross-fitting.\n",
    "In case of two-way clustering, we assume $N$ clusters in the first dimension and $M$ clusters in the second dimension.\n",
    "\n",
    "For $K$-fold cross-fitting, [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815) proposed to randomly partition $[N]:=\\{1,\\ldots,N\\}$ into $K$ subsets $\\{I_1, \\ldots, I_K\\}$ and $[M]:=\\{1,\\ldots,N\\}$ into $K$ subsets $\\{J_1, \\ldots, J_K\\}$.\n",
    "Effectively, one then considers $K^2$ folds.\n",
    "Basically for each $(k, \\ell) \\in \\{1, \\ldots, K\\} \\times \\{1, \\ldots, K\\}$, the nuisance functions are estimated for all double-indexed observations in $([N]\\setminus I_K) \\times ([M]\\setminus J_\\ell)$, i.e.,\n",
    "$$\n",
    "\\hat{\\eta}_{k\\ell} = \\hat{\\eta}\\left((W_{ij})_{(i,j)\\in ([N]\\setminus I_K) \\times ([M]\\setminus J_\\ell)}\\right)\n",
    "$$\n",
    "The causal parameter is then estimated as usual by solving a moment condition with a Neyman orthogonal score function.\n",
    "For two-way cluster robust double machine learning with algorithm [DML2](https://docs.doubleml.org/stable/guide/algorithms.html#algorithm-dml2) this results in solving\n",
    "$$\n",
    "\\frac{1}{K^2} \\sum_{k=1}^{K} \\sum_{\\ell=1}^{K} \\frac{1}{|I_k| |J_\\ell|} \\sum_{(i,j) \\in I_K \\times J_\\ell}\n",
    "\\psi(W_{ij}, \\tilde{\\theta}_0, \\hat{\\eta}_{k\\ell}) = 0\n",
    "$$\n",
    "for $\\tilde{\\theta}_0$.\n",
    "Here $|I_k|$ denotes the cardinality, i.e., the number of clusters in the $k$-th fold for the first cluster variable."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "CwgzBXvQ1_am"
   },
   "source": [
    "We can visualize the sample splitting of the $N \\cdot M = 625$ observations into $K \\cdot K = 9$ folds. The following heat map illustrates the partitioned data set that is split into $K=9$ folds. The horizontal axis corresponds to the fold indices and the vertical axis to the indices of the observations. A blue field indicates that the observation $i$ is used for fitting the nuisance part, red indicates that the fold is used for prediction generation and white means that an observation is left out from the sample splitting. \n",
    "\n",
    "For example, the first observation as displayed on the very bottom of the figure (using Python indexing starting at `0`) is used for training of the nuisance parts in the first (`0`), the third (`2`), fourth (`3`) and sixth (`5`) fold and used for generation of the predictions in fold eight (`7`). At the same time the observation is left out from the sample splitting procedure in folds two (`1`), five (`4`), seven (`6`) and nine (`8`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "2eU-ukl81_ar"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x1200 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# The function plt_smpls is defined at the end of the Notebook\n",
    "plt_smpls(dml_pliv_obj.smpls[0], dml_pliv_obj._n_folds_per_cluster)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "WbylaFQx1_ay"
   },
   "source": [
    "If we visualize the sample splitting in terms of the cluster variables, the partitioning of the data into $9$ folds $I_k \\times J_\\ell$ becomes clear.\n",
    "The identifiers for the first cluster variable $[N]:=\\{1,\\ldots,N\\}$ have been randomly partioned into $K=3$ folds denoted by $\\{I_1, I_2, I_3\\}$ and the identifiers for the second cluster variable $[M]:=\\{1,\\ldots,M\\}$ have also been randomly partioned into $K=3$ folds denoted by $\\{J_1, J_2, J_3\\}$.\n",
    "By considering every combination $I_k \\times J_\\ell$ for $1 \\leq k, \\ell \\leq K = 3$ we effectively base the cross-fitting on $9$ folds.\n",
    "\n",
    "We now want to focus on the top-left sub-plot showing the partitioning of the cluster data for the first fold.\n",
    "The $x$-axis corresponds to the first cluster variable and the $y$-axis to the second cluster variable.\n",
    "Observations with cluster variables $(i,j) \\in I_K \\times J_\\ell$ are used for estimation of the target parameter $\\tilde{\\theta}_0$ by solving a Neyman orthogonal score function.\n",
    "For estimation of the nuisance function, we only use observation where neither the first cluster variable is in $I_K$ nor the second cluster variable is in $J_\\ell$, i.e., we use observations indexed by $(i,j)\\in ([N]\\setminus I_K) \\times ([M]\\setminus J_\\ell)$ to estimate the nuisance functions\n",
    "$$\n",
    "\\hat{\\eta}_{k\\ell} = \\hat{\\eta}\\left((W_{ij})_{(i,j)\\in ([N]\\setminus I_K) \\times ([M]\\setminus J_\\ell)}\\right).\n",
    "$$\n",
    "This way we guarantee that there are never observations from the same cluster (first and/or second cluster dimension) in the sample for the nuisance function estimation (blue) and at the same time in the sample for solving the score function (red). As a result of this special sample splitting proposed by [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815), the observations in the score (red) and nuisance (blue) sample can be considered independent and the standard cross-fitting approach for double machine learning can be applied."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "KfI-DvHV1_a1",
    "tags": [
     "nbsphinx-thumbnail"
    ]
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x1200 with 18 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# The function plt_smpls_cluster is defined at the end of the Notebook\n",
    "plt_smpls_cluster(dml_pliv_obj.smpls_cluster[0], dml_pliv_obj._n_folds_per_cluster)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "wS8OuXQG1_a5"
   },
   "source": [
    "### Cluster Robust Standard Errors\n",
    "In the abstract base class `DoubleML` the estimation of cluster robust standard errors is implemented for all supported double machine learning models.\n",
    "It is based on the assumption of a linear Neyman orthogonal score function.\n",
    "We use the notation $n \\wedge m := \\min\\{n,m\\}$.\n",
    "For the the asymptotic variance of\n",
    "$\\sqrt{\\underline{C}}(\\tilde{\\theta_0} - \\theta_0)$ with\n",
    "$\\underline{C} := N \\wedge M$\n",
    "[Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815) then propose the following estimator\n",
    "$$\n",
    "\\hat{\\sigma}^2 = \\hat{J}^{-1} \\hat{\\Gamma} \\hat{J}^{-1}\n",
    "$$\n",
    "where\n",
    "$$\n",
    "\\begin{aligned}\n",
    "\\hat{\\Gamma} = \\frac{1}{K^2} \\sum_{(k, \\ell) \\in[K]^2}\n",
    "\\Bigg[ \\frac{|I_k| \\wedge |J_\\ell|}{(|I_k||J_\\ell|)^2}\n",
    "\\bigg(&\\sum_{i \\in I_k} \\sum_{j \\in J_\\ell} \\sum_{j' \\in J_\\ell}\n",
    "\\psi(W_{ij}; \\tilde{\\theta}, \\hat{\\eta}_{k \\ell}) \\psi(W_{ij'}; \\tilde{\\theta}_0, \\hat{\\eta}_{k \\ell}) \\\\\n",
    "&+ \\sum_{i \\in I_k} \\sum_{i' \\in I_k} \\sum_{j \\in J_\\ell}\n",
    "\\psi(W_{ij}; \\tilde{\\theta}, \\hat{\\eta}_{k \\ell}) \\psi(W_{i'j}; \\tilde{\\theta}_0, \\hat{\\eta}_{k \\ell})\n",
    "\\bigg)\n",
    "\\Bigg]\n",
    "\\end{aligned}$$\n",
    "and\n",
    "$$\n",
    "\\begin{aligned}\n",
    "\\hat{J} = \\frac{1}{K^2} \\sum_{(k, \\ell) \\in[K]^2} \\frac{1}{|I_k||J_\\ell|}\n",
    "\\sum_{i \\in I_k} \\sum_{j \\in J_\\ell}\n",
    "\\psi_a(W_{ij}; \\tilde{\\theta}_0, \\hat{\\eta}_{k \\ell}).\n",
    "\\end{aligned}\n",
    "$$\n",
    "A $(1-\\alpha)$ confidence interval is then given by ([Chiang et al. 2021](https://doi.org/10.1080/07350015.2021.1895815))\n",
    "$$\\begin{aligned}\n",
    "\\left[\n",
    "\\tilde{\\theta} \\pm \\Phi^{-1}(1-\\alpha/2) \\sqrt{\\hat{\\sigma}^2 / \\underline{C}}\n",
    "\\right]\n",
    "\\end{aligned}\n",
    "$$\n",
    "with $\\underline{C} = N \\wedge M$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "ZcNPzPnC1_a9"
   },
   "outputs": [
    {
     "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>coef</th>\n",
       "      <th>std err</th>\n",
       "      <th>t</th>\n",
       "      <th>P&gt;|t|</th>\n",
       "      <th>2.5 %</th>\n",
       "      <th>97.5 %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>D</th>\n",
       "      <td>1.135871</td>\n",
       "      <td>0.118601</td>\n",
       "      <td>9.577271</td>\n",
       "      <td>9.964261e-22</td>\n",
       "      <td>0.903418</td>\n",
       "      <td>1.368324</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       coef   std err         t         P>|t|     2.5 %    97.5 %\n",
       "D  1.135871  0.118601  9.577271  9.964261e-22  0.903418  1.368324"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Estimate the PLIV model with cluster robust double machine learning\n",
    "dml_pliv_obj.fit()\n",
    "dml_pliv_obj.summary"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "XNRVgYrF1_bA"
   },
   "source": [
    "## (One-Way) Cluster Robust Double Machine Learing\n",
    "\n",
    "We again use the PLIV data generating process described in Section 4.1 of [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815).\n",
    "To obtain one-way clustered data, we set the following weights to zero\n",
    "$$\n",
    "\\omega_2^X = \\omega_2^\\varepsilon = \\omega_2^v = \\omega_2^V = 0.\n",
    "$$\n",
    "Again we can simulate this data with [make_pliv_multiway_cluster_CKMS2021()](https://docs.doubleml.org/stable/api/generated/doubleml.datasets.make_pliv_multiway_cluster_CKMS2021.html#doubleml.datasets.make_pliv_multiway_cluster_CKMS2021). To prepare the data-backend for one-way clustering, we only have to alter the `cluster_cols` to be `'cluster_var_i'`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "EvJ-wHSV1_bD"
   },
   "outputs": [],
   "source": [
    "obj_dml_data = make_pliv_multiway_cluster_CKMS2021(N, M, dim_X,\n",
    "                                                   omega_X=np.array([0.25, 0]),\n",
    "                                                   omega_epsilon=np.array([0.25, 0]),\n",
    "                                                   omega_v=np.array([0.25, 0]),\n",
    "                                                   omega_V=np.array([0.25, 0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "UIMLXmD51_bF"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== DoubleMLClusterData Object ==================\n",
      "\n",
      "------------------ Data summary      ------------------\n",
      "Outcome variable: Y\n",
      "Treatment variable(s): ['D']\n",
      "Cluster variable(s): ['cluster_var_i']\n",
      "Covariates: ['X1', 'X2', 'X3', 'X4', 'X5', 'X6', 'X7', 'X8', 'X9', 'X10', 'X11', 'X12', 'X13', 'X14', 'X15', 'X16', 'X17', 'X18', 'X19', 'X20', 'X21', 'X22', 'X23', 'X24', 'X25', 'X26', 'X27', 'X28', 'X29', 'X30', 'X31', 'X32', 'X33', 'X34', 'X35', 'X36', 'X37', 'X38', 'X39', 'X40', 'X41', 'X42', 'X43', 'X44', 'X45', 'X46', 'X47', 'X48', 'X49', 'X50', 'X51', 'X52', 'X53', 'X54', 'X55', 'X56', 'X57', 'X58', 'X59', 'X60', 'X61', 'X62', 'X63', 'X64', 'X65', 'X66', 'X67', 'X68', 'X69', 'X70', 'X71', 'X72', 'X73', 'X74', 'X75', 'X76', 'X77', 'X78', 'X79', 'X80', 'X81', 'X82', 'X83', 'X84', 'X85', 'X86', 'X87', 'X88', 'X89', 'X90', 'X91', 'X92', 'X93', 'X94', 'X95', 'X96', 'X97', 'X98', 'X99', 'X100']\n",
      "Instrument variable(s): ['Z']\n",
      "No. Observations: 625\n",
      "\n",
      "------------------ DataFrame info    ------------------\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 625 entries, 0 to 624\n",
      "Columns: 105 entries, cluster_var_i to Z\n",
      "dtypes: float64(103), int64(2)\n",
      "memory usage: 512.8 KB\n",
      "\n"
     ]
    }
   ],
   "source": [
    "obj_dml_data.cluster_cols = 'cluster_var_i'\n",
    "print(obj_dml_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "Jb0zciH41_bJ"
   },
   "outputs": [],
   "source": [
    "# Set machine learning methods for l, m & r\n",
    "learner = LassoCV()\n",
    "ml_l = clone(learner)\n",
    "ml_m = clone(learner)\n",
    "ml_r = clone(learner)\n",
    "\n",
    "# initialize the DoubleMLPLIV object\n",
    "dml_pliv_obj = DoubleMLPLIV(obj_dml_data,\n",
    "                            ml_l, ml_m, ml_r,\n",
    "                            n_folds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "DcQHiCFG1_bN"
   },
   "outputs": [
    {
     "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>coef</th>\n",
       "      <th>std err</th>\n",
       "      <th>t</th>\n",
       "      <th>P&gt;|t|</th>\n",
       "      <th>2.5 %</th>\n",
       "      <th>97.5 %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>D</th>\n",
       "      <td>0.951532</td>\n",
       "      <td>0.047954</td>\n",
       "      <td>19.842625</td>\n",
       "      <td>1.276189e-87</td>\n",
       "      <td>0.857544</td>\n",
       "      <td>1.04552</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       coef   std err          t         P>|t|     2.5 %   97.5 %\n",
       "D  0.951532  0.047954  19.842625  1.276189e-87  0.857544  1.04552"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dml_pliv_obj.fit()\n",
    "dml_pliv_obj.summary"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "BViUo9UW1_bZ"
   },
   "source": [
    "## Real-Data Application\n",
    "As a real-data application we revist the consumer demand example from [Chiang et al. (2021)](https://doi.org/10.1080/07350015.2021.1895815).\n",
    "The U.S. automobile data of [Berry, Levinsohn, and Pakes (1995)](https://doi.org/10.2307/2171802) is obtained from the `R` package [hdm](https://cran.r-project.org/web/packages/hdm/index.html). In this example, we consider different specifications for the cluster dimensions."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "A7p7eaeO1_bb"
   },
   "source": [
    "### Load and Process Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "SM30YzZo1_bh"
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from rpy2.robjects.packages import PackageData\n",
    "from rpy2.robjects import pandas2ri, default_converter, conversion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "kWO-X1aZ1_bj"
   },
   "outputs": [],
   "source": [
    "r_df = PackageData('hdm').fetch('BLP')['BLP'][0]\n",
    "with conversion.localconverter(default_converter + pandas2ri.converter):\n",
    "    blp_data = conversion.rpy2py(r_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "Kzu6GG2i1_bo"
   },
   "outputs": [
    {
     "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>model.name</th>\n",
       "      <th>model.id</th>\n",
       "      <th>firm.id</th>\n",
       "      <th>cdid</th>\n",
       "      <th>id</th>\n",
       "      <th>price</th>\n",
       "      <th>mpd</th>\n",
       "      <th>air</th>\n",
       "      <th>mpg</th>\n",
       "      <th>space</th>\n",
       "      <th>hpwt</th>\n",
       "      <th>trend</th>\n",
       "      <th>share</th>\n",
       "      <th>outshr</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AMGREM</td>\n",
       "      <td>23.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>-6.825617</td>\n",
       "      <td>1.888146</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.697</td>\n",
       "      <td>1.150200</td>\n",
       "      <td>0.528997</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001051</td>\n",
       "      <td>0.880106</td>\n",
       "      <td>0.820366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AMHORN</td>\n",
       "      <td>24.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>-6.245370</td>\n",
       "      <td>1.935989</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.740</td>\n",
       "      <td>1.278000</td>\n",
       "      <td>0.494324</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000670</td>\n",
       "      <td>0.880106</td>\n",
       "      <td>0.369981</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AMJAVL</td>\n",
       "      <td>25.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>132.0</td>\n",
       "      <td>-4.652778</td>\n",
       "      <td>1.716799</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.543</td>\n",
       "      <td>1.459200</td>\n",
       "      <td>0.467613</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000341</td>\n",
       "      <td>0.880106</td>\n",
       "      <td>-0.306915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AMMATA</td>\n",
       "      <td>26.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>134.0</td>\n",
       "      <td>-4.921913</td>\n",
       "      <td>1.687871</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.517</td>\n",
       "      <td>1.606800</td>\n",
       "      <td>0.426540</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000522</td>\n",
       "      <td>0.880106</td>\n",
       "      <td>0.120721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>AMAMBS</td>\n",
       "      <td>18.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>-2.833024</td>\n",
       "      <td>1.504286</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.352</td>\n",
       "      <td>1.645800</td>\n",
       "      <td>0.452489</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000442</td>\n",
       "      <td>0.880106</td>\n",
       "      <td>-0.045144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2213</th>\n",
       "      <td>VV740</td>\n",
       "      <td>526.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>5584.0</td>\n",
       "      <td>4.378596</td>\n",
       "      <td>2.639135</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.100</td>\n",
       "      <td>1.305612</td>\n",
       "      <td>0.385917</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.000488</td>\n",
       "      <td>0.907801</td>\n",
       "      <td>0.022915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2214</th>\n",
       "      <td>VV760G</td>\n",
       "      <td>529.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>5585.0</td>\n",
       "      <td>14.225574</td>\n",
       "      <td>2.136442</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.700</td>\n",
       "      <td>1.305612</td>\n",
       "      <td>0.435967</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.000091</td>\n",
       "      <td>0.907801</td>\n",
       "      <td>-1.652350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2215</th>\n",
       "      <td>YGGVPL</td>\n",
       "      <td>555.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>5589.0</td>\n",
       "      <td>-8.368152</td>\n",
       "      <td>3.518846</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.800</td>\n",
       "      <td>0.843730</td>\n",
       "      <td>0.358289</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.000067</td>\n",
       "      <td>0.907801</td>\n",
       "      <td>-1.957375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2216</th>\n",
       "      <td>PS911C</td>\n",
       "      <td>425.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>5590.0</td>\n",
       "      <td>32.997571</td>\n",
       "      <td>3.016154</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.400</td>\n",
       "      <td>1.093950</td>\n",
       "      <td>0.814913</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.000039</td>\n",
       "      <td>0.907801</td>\n",
       "      <td>-2.512519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2217</th>\n",
       "      <td>PS944</td>\n",
       "      <td>438.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>5592.0</td>\n",
       "      <td>20.296729</td>\n",
       "      <td>3.267500</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.600</td>\n",
       "      <td>1.153587</td>\n",
       "      <td>0.693796</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.000025</td>\n",
       "      <td>0.907801</td>\n",
       "      <td>-2.953683</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2217 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     model.name  model.id  firm.id  cdid      id      price       mpd  air  \\\n",
       "1        AMGREM      23.0     15.0   1.0   129.0  -6.825617  1.888146  0.0   \n",
       "2        AMHORN      24.0     15.0   1.0   130.0  -6.245370  1.935989  0.0   \n",
       "3        AMJAVL      25.0     15.0   1.0   132.0  -4.652778  1.716799  0.0   \n",
       "4        AMMATA      26.0     15.0   1.0   134.0  -4.921913  1.687871  0.0   \n",
       "5        AMAMBS      18.0     15.0   1.0   136.0  -2.833024  1.504286  0.0   \n",
       "...         ...       ...      ...   ...     ...        ...       ...  ...   \n",
       "2213     VV740      526.0      6.0  20.0  5584.0   4.378596  2.639135  1.0   \n",
       "2214     VV760G     529.0      6.0  20.0  5585.0  14.225574  2.136442  1.0   \n",
       "2215     YGGVPL     555.0     23.0  20.0  5589.0  -8.368152  3.518846  0.0   \n",
       "2216     PS911C     425.0     12.0  20.0  5590.0  32.997571  3.016154  1.0   \n",
       "2217     PS944      438.0     12.0  20.0  5592.0  20.296729  3.267500  1.0   \n",
       "\n",
       "        mpg     space      hpwt  trend     share    outshr         y  \n",
       "1     1.697  1.150200  0.528997    0.0  0.001051  0.880106  0.820366  \n",
       "2     1.740  1.278000  0.494324    0.0  0.000670  0.880106  0.369981  \n",
       "3     1.543  1.459200  0.467613    0.0  0.000341  0.880106 -0.306915  \n",
       "4     1.517  1.606800  0.426540    0.0  0.000522  0.880106  0.120721  \n",
       "5     1.352  1.645800  0.452489    0.0  0.000442  0.880106 -0.045144  \n",
       "...     ...       ...       ...    ...       ...       ...       ...  \n",
       "2213  2.100  1.305612  0.385917   19.0  0.000488  0.907801  0.022915  \n",
       "2214  1.700  1.305612  0.435967   19.0  0.000091  0.907801 -1.652350  \n",
       "2215  2.800  0.843730  0.358289   19.0  0.000067  0.907801 -1.957375  \n",
       "2216  2.400  1.093950  0.814913   19.0  0.000039  0.907801 -2.512519  \n",
       "2217  2.600  1.153587  0.693796   19.0  0.000025  0.907801 -2.953683  \n",
       "\n",
       "[2217 rows x 15 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_cols = ['hpwt', 'air', 'mpd', 'space']\n",
    "blp_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "5WIE4GPS1_bq"
   },
   "outputs": [],
   "source": [
    "def construct_iv(blp_data, x_cols=['hpwt', 'air', 'mpd', 'space']):\n",
    "    n = blp_data.shape[0]\n",
    "    p = len(x_cols)\n",
    "    \n",
    "    firmid = blp_data['firm.id'].values\n",
    "    cdid = blp_data['cdid'].values\n",
    "    id_var = blp_data['id'].values\n",
    "    X = blp_data[x_cols]\n",
    "    \n",
    "    sum_other = pd.DataFrame(columns=['sum.other.' + var for var in x_cols],\n",
    "                             index=blp_data.index)\n",
    "    sum_rival = pd.DataFrame(columns=['sum.rival.' + var for var in x_cols],\n",
    "                             index=blp_data.index)\n",
    "    \n",
    "    for i in range(n):\n",
    "        other_ind = (firmid == firmid[i]) & (cdid == cdid[i]) & (id_var != id_var[i])\n",
    "        rival_ind = (firmid != firmid[i]) & (cdid == cdid[i])\n",
    "        for j in range(p):\n",
    "            sum_other.iloc[i, j] = X.iloc[:,j][other_ind].sum()\n",
    "            sum_rival.iloc[i, j] = X.iloc[:,j][rival_ind].sum()\n",
    "    \n",
    "    return pd.concat((sum_other, sum_rival), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "id": "2iTzEmsT1_br"
   },
   "outputs": [],
   "source": [
    "iv_vars = construct_iv(blp_data, x_cols=['hpwt', 'air', 'mpd', 'space'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "vjJEPgd61_bs"
   },
   "outputs": [],
   "source": [
    "poly = PolynomialFeatures(degree=3, include_bias=False)\n",
    "data_transf = poly.fit_transform(blp_data[x_cols])\n",
    "x_cols_poly = poly.get_feature_names_out(x_cols)\n",
    "data_transf = pd.DataFrame(data_transf, columns=x_cols_poly)\n",
    "data_transf.index = blp_data.index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "id": "VzG84J7K1_b3",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['air',\n",
       " 'air mpd',\n",
       " 'air mpd^2',\n",
       " 'air space',\n",
       " 'air space^2',\n",
       " 'air^2',\n",
       " 'air^2 mpd',\n",
       " 'air^2 space',\n",
       " 'air^3',\n",
       " 'hpwt',\n",
       " 'hpwt air',\n",
       " 'hpwt air^2',\n",
       " 'hpwt mpd',\n",
       " 'hpwt mpd^2',\n",
       " 'hpwt space',\n",
       " 'hpwt space^2',\n",
       " 'hpwt^2',\n",
       " 'hpwt^2 air',\n",
       " 'hpwt^2 mpd',\n",
       " 'hpwt^2 space',\n",
       " 'hpwt^3',\n",
       " 'mpd',\n",
       " 'mpd space',\n",
       " 'mpd space^2',\n",
       " 'mpd^2',\n",
       " 'mpd^2 space',\n",
       " 'mpd^3',\n",
       " 'space',\n",
       " 'space^2',\n",
       " 'space^3']"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_cols_chiang = list(np.setdiff1d(data_transf.columns,\n",
    "                                    ['hpwt air mpd', 'hpwt air space',\n",
    "                                     'hpwt mpd space', 'air mpd space']))\n",
    "sel_cols_chiang"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "M8IeK-0A1_b6"
   },
   "outputs": [],
   "source": [
    "blp_data['log_p'] = np.log(blp_data['price'] + 11.761)\n",
    "\n",
    "y_col = 'y'\n",
    "d_col = 'log_p'\n",
    "cluster_cols = ['model.id', 'cdid']\n",
    "all_z_cols = ['sum.other.hpwt', 'sum.other.mpd', 'sum.other.space']\n",
    "z_col = all_z_cols[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "7eVrppKH1_b7"
   },
   "outputs": [],
   "source": [
    "dml_df = pd.concat((blp_data[[y_col] + [d_col] + cluster_cols],\n",
    "                    data_transf[sel_cols_chiang],\n",
    "                    iv_vars[all_z_cols]),\n",
    "                   axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "id": "ezGVW5Gl1_b8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2217, 37)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dml_df.shape"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "05nt4NKc1_b9"
   },
   "source": [
    "### Initialize `DoubleMLClusterData` object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "id": "ik3fUxCn1_cH"
   },
   "outputs": [],
   "source": [
    "dml_data = DoubleMLClusterData(dml_df,\n",
    "                               y_col=y_col,\n",
    "                               d_cols=d_col,\n",
    "                               z_cols=z_col,\n",
    "                               cluster_cols=cluster_cols,\n",
    "                               x_cols=sel_cols_chiang)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "id": "7Y4wQOsC1_cJ"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== DoubleMLClusterData Object ==================\n",
      "\n",
      "------------------ Data summary      ------------------\n",
      "Outcome variable: y\n",
      "Treatment variable(s): ['log_p']\n",
      "Cluster variable(s): ['model.id', 'cdid']\n",
      "Covariates: ['air', 'air mpd', 'air mpd^2', 'air space', 'air space^2', 'air^2', 'air^2 mpd', 'air^2 space', 'air^3', 'hpwt', 'hpwt air', 'hpwt air^2', 'hpwt mpd', 'hpwt mpd^2', 'hpwt space', 'hpwt space^2', 'hpwt^2', 'hpwt^2 air', 'hpwt^2 mpd', 'hpwt^2 space', 'hpwt^3', 'mpd', 'mpd space', 'mpd space^2', 'mpd^2', 'mpd^2 space', 'mpd^3', 'space', 'space^2', 'space^3']\n",
      "Instrument variable(s): ['sum.other.hpwt']\n",
      "No. Observations: 2217\n",
      "\n",
      "------------------ DataFrame info    ------------------\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 2217 entries, 1 to 2217\n",
      "Columns: 37 entries, y to sum.other.space\n",
      "dtypes: float64(34), object(3)\n",
      "memory usage: 658.2+ KB\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(dml_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "id": "LGgk1Yfy1_cK"
   },
   "outputs": [],
   "source": [
    "learner = LassoCV(max_iter=50000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "id": "GssTYt2i1_cM"
   },
   "outputs": [],
   "source": [
    "res_df = pd.DataFrame()\n",
    "n_rep = 10"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "wsagdd4V1_cN"
   },
   "source": [
    "### Two-Way Clustering with Respect to Product and Market"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "id": "Dd2InB0w1_cO"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\svenk\\AppData\\Roaming\\Python\\Python310\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:620: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Duality gap: 3.5154789948092002, tolerance: 1.1187339840850312\n",
      "  model = cd_fast.enet_coordinate_descent_gram(\n"
     ]
    }
   ],
   "source": [
    "dml_data.z_cols = z_col\n",
    "dml_data.cluster_cols = ['model.id', 'cdid']\n",
    "dml_pliv = DoubleMLPLIV(dml_data,\n",
    "                        clone(learner), clone(learner), clone(learner),\n",
    "                        n_folds=2, n_rep=n_rep)\n",
    "dml_pliv.fit()\n",
    "res = dml_pliv.summary.reset_index(drop=True)\n",
    "res['z_col'] = dml_data.z_cols[0]\n",
    "res['clustering'] = 'two-way'\n",
    "res_df = pd.concat([res_df, res]).reset_index(drop=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "n3RhZ4fH1_cS"
   },
   "source": [
    "### One-Way Clustering with Respect to the Product"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "id": "dqJxk4S61_cV"
   },
   "outputs": [],
   "source": [
    "dml_data.z_cols = z_col\n",
    "dml_data.cluster_cols = 'model.id'\n",
    "dml_pliv = DoubleMLPLIV(dml_data,\n",
    "                        clone(learner), clone(learner), clone(learner),\n",
    "                        n_folds=4, n_rep=n_rep)\n",
    "dml_pliv.fit()\n",
    "res = dml_pliv.summary.reset_index(drop=True)\n",
    "res['z_col'] = dml_data.z_cols[0]\n",
    "res['clustering'] = 'one-way-product'\n",
    "res_df = pd.concat([res_df, res]).reset_index(drop=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "ODt4TiKL1_cX"
   },
   "source": [
    "### One-Way Clustering with Respect to the Market"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "id": "dl2CTDBu1_cY"
   },
   "outputs": [],
   "source": [
    "dml_data.z_cols = z_col\n",
    "dml_data.cluster_cols = 'cdid'\n",
    "dml_pliv = DoubleMLPLIV(dml_data,\n",
    "                        clone(learner), clone(learner), clone(learner),\n",
    "                        n_folds=4, n_rep=n_rep)\n",
    "dml_pliv.fit()\n",
    "res = dml_pliv.summary.reset_index(drop=True)\n",
    "res['z_col'] = dml_data.z_cols[0]\n",
    "res['clustering'] = 'one-way-market'\n",
    "res_df = pd.concat([res_df, res]).reset_index(drop=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "96RUevUV1_cZ"
   },
   "source": [
    "### No Clustering / Zero-Way Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "id": "wkKU9Fxx1_cc"
   },
   "outputs": [],
   "source": [
    "dml_data = DoubleMLData(dml_df,\n",
    "                        y_col=y_col,\n",
    "                        d_cols=d_col,\n",
    "                        z_cols=z_col,\n",
    "                        x_cols=sel_cols_chiang)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "id": "xxCoX1CP1_cd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== DoubleMLData Object ==================\n",
      "\n",
      "------------------ Data summary      ------------------\n",
      "Outcome variable: y\n",
      "Treatment variable(s): ['log_p']\n",
      "Covariates: ['air', 'air mpd', 'air mpd^2', 'air space', 'air space^2', 'air^2', 'air^2 mpd', 'air^2 space', 'air^3', 'hpwt', 'hpwt air', 'hpwt air^2', 'hpwt mpd', 'hpwt mpd^2', 'hpwt space', 'hpwt space^2', 'hpwt^2', 'hpwt^2 air', 'hpwt^2 mpd', 'hpwt^2 space', 'hpwt^3', 'mpd', 'mpd space', 'mpd space^2', 'mpd^2', 'mpd^2 space', 'mpd^3', 'space', 'space^2', 'space^3']\n",
      "Instrument variable(s): ['sum.other.hpwt']\n",
      "No. Observations: 2217\n",
      "\n",
      "------------------ DataFrame info    ------------------\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 2217 entries, 1 to 2217\n",
      "Columns: 37 entries, y to sum.other.space\n",
      "dtypes: float64(34), object(3)\n",
      "memory usage: 658.2+ KB\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(dml_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "id": "WY3iQ5aO1_cq"
   },
   "outputs": [],
   "source": [
    "dml_data.z_cols = z_col\n",
    "dml_pliv = DoubleMLPLIV(dml_data,\n",
    "                        clone(learner), clone(learner), clone(learner),\n",
    "                        n_folds=4, n_rep=n_rep)\n",
    "dml_pliv.fit()\n",
    "res = dml_pliv.summary.reset_index(drop=True)\n",
    "res['z_col'] = dml_data.z_cols[0]\n",
    "res['clustering'] = 'zero-way'\n",
    "res_df = pd.concat([res_df, res]).reset_index(drop=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "Hmc3x1n-1_cr"
   },
   "source": [
    "### Application Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "id": "PO-RMWTe1_cs"
   },
   "outputs": [
    {
     "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>coef</th>\n",
       "      <th>std err</th>\n",
       "      <th>t</th>\n",
       "      <th>P&gt;|t|</th>\n",
       "      <th>2.5 %</th>\n",
       "      <th>97.5 %</th>\n",
       "      <th>z_col</th>\n",
       "      <th>clustering</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-5.450152</td>\n",
       "      <td>1.262621</td>\n",
       "      <td>-4.316540</td>\n",
       "      <td>1.584942e-05</td>\n",
       "      <td>-7.924843</td>\n",
       "      <td>-2.975461</td>\n",
       "      <td>sum.other.hpwt</td>\n",
       "      <td>two-way</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-5.757819</td>\n",
       "      <td>0.928947</td>\n",
       "      <td>-6.198218</td>\n",
       "      <td>5.710586e-10</td>\n",
       "      <td>-7.578523</td>\n",
       "      <td>-3.937116</td>\n",
       "      <td>sum.other.hpwt</td>\n",
       "      <td>one-way-product</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-5.811825</td>\n",
       "      <td>0.784483</td>\n",
       "      <td>-7.408479</td>\n",
       "      <td>1.277561e-13</td>\n",
       "      <td>-7.349383</td>\n",
       "      <td>-4.274267</td>\n",
       "      <td>sum.other.hpwt</td>\n",
       "      <td>one-way-market</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-5.754870</td>\n",
       "      <td>0.464284</td>\n",
       "      <td>-12.395136</td>\n",
       "      <td>2.776728e-35</td>\n",
       "      <td>-6.664850</td>\n",
       "      <td>-4.844889</td>\n",
       "      <td>sum.other.hpwt</td>\n",
       "      <td>zero-way</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       coef   std err          t         P>|t|     2.5 %    97.5 %  \\\n",
       "0 -5.450152  1.262621  -4.316540  1.584942e-05 -7.924843 -2.975461   \n",
       "1 -5.757819  0.928947  -6.198218  5.710586e-10 -7.578523 -3.937116   \n",
       "2 -5.811825  0.784483  -7.408479  1.277561e-13 -7.349383 -4.274267   \n",
       "3 -5.754870  0.464284 -12.395136  2.776728e-35 -6.664850 -4.844889   \n",
       "\n",
       "            z_col       clustering  \n",
       "0  sum.other.hpwt          two-way  \n",
       "1  sum.other.hpwt  one-way-product  \n",
       "2  sum.other.hpwt   one-way-market  \n",
       "3  sum.other.hpwt         zero-way  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_df"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "NhIE_t-A1_ct"
   },
   "source": [
    "## References\n",
    "Berry, S., Levinsohn, J., and Pakes, A. (1995), Automobile Prices in Market\n",
    "Equilibrium, Econometrica: Journal of the Econometric Society, 63, 841-890, doi: [10.2307/2171802](https://doi.org/10.2307/2171802).\n",
    "\n",
    "Cameron, A. C., Gelbach, J. B. and Miller, D. L. (2011), Robust Inference with Multiway Clustering, Journal of Business & Economic Statistics, 29:2, 238-249, doi: [10.1198/jbes.2010.07136](https://doi.org/10.1198/jbes.2010.07136).\n",
    "\n",
    "Chernozhukov, V., Chetverikov, D., Demirer, M., Duflo, E., Hansen, C., Newey, W. and Robins, J. (2018), Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal, 21: C1-C68, doi: [10.1111/ectj.12097](https://doi.org/10.1111/ectj.12097).\n",
    "\n",
    "Chiang, H. D., Kato K., Ma, Y. and Sasaki, Y. (2021), Multiway Cluster Robust Double/Debiased Machine Learning, Journal of Business & Economic Statistics, doi: [10.1080/07350015.2021.1895815](https://doi.org/10.1080/07350015.2021.1895815), arXiv: [1909.03489](https://arxiv.org/abs/1909.03489)."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define Helper Functions for Plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "#discrete color scheme\n",
    "x = sns.color_palette(\"RdBu_r\", 7)\n",
    "cMap = ListedColormap([x[0], x[3], x[6]])\n",
    "plt.rcParams['figure.figsize'] = 15, 12\n",
    "sns.set(font_scale=1.3)\n",
    "\n",
    "def plt_smpls(smpls, n_folds_per_cluster):\n",
    "    df = pd.DataFrame(np.zeros([N*M, n_folds_per_cluster*n_folds_per_cluster]))\n",
    "    for i_split, this_split_ind in enumerate(smpls):\n",
    "        df.loc[this_split_ind[0], i_split] = -1.\n",
    "        df.loc[this_split_ind[1], i_split] = 1.\n",
    "\n",
    "    ax = sns.heatmap(df, cmap=cMap);\n",
    "    ax.invert_yaxis();\n",
    "    ax.set_ylim([0, N*M]);\n",
    "    ax.set_xlabel('Fold')\n",
    "    ax.set_ylabel('Observation')\n",
    "    colorbar = ax.collections[0].colorbar\n",
    "    colorbar.set_ticks([-0.667, 0, 0.667])\n",
    "    colorbar.set_ticklabels(['Nuisance', '', 'Score'])\n",
    "\n",
    "def plt_smpls_cluster(smpls_cluster, n_folds_per_cluster):\n",
    "    for i_split in range(len(smpls_cluster)):\n",
    "        plt.subplot(n_folds_per_cluster, n_folds_per_cluster, i_split + 1)\n",
    "        df = pd.DataFrame(np.zeros([N*M, 1]),\n",
    "                      index = pd.MultiIndex.from_product([range(N), range(M)]),\n",
    "                      columns=['value'])\n",
    "\n",
    "        df.loc[pd.MultiIndex.from_product(smpls_cluster[i_split][0]), :] = -1.\n",
    "        df.loc[pd.MultiIndex.from_product(smpls_cluster[i_split][1]), :] = 1.\n",
    "\n",
    "        df_wide = df.reset_index().pivot(index=\"level_0\", columns=\"level_1\", values=\"value\")\n",
    "        df_wide.index.name=''\n",
    "        df_wide.columns.name=''\n",
    "\n",
    "        ax = sns.heatmap(df_wide, cmap=cMap);\n",
    "        ax.invert_yaxis();\n",
    "        ax.set_ylim([0, M]);\n",
    "        colorbar = ax.collections[0].colorbar\n",
    "        colorbar.set_ticks([-0.667, 0, 0.667])\n",
    "\n",
    "        l = i_split % n_folds_per_cluster + 1\n",
    "        k = np.floor_divide(i_split, n_folds_per_cluster) + 1\n",
    "        title = f'Nuisance: $I_{{{k}}}^C \\\\times J_{{{l}}}^C$; Score: $I_{{{k}}} \\\\times J_{{{l}}}$'\n",
    "        ax.set_title(title)\n",
    "        if l == n_folds_per_cluster:\n",
    "            colorbar.set_ticklabels(['Nuisance', '', 'Score'])\n",
    "        else:\n",
    "            colorbar.set_ticklabels(['', '', ''])\n",
    "        if l == 1:\n",
    "            ax.set_ylabel('First Cluster Variable $k$')\n",
    "        if k == 3:\n",
    "            ax.set_xlabel('Second Cluster Variable $\\ell$')\n",
    "    plt.tight_layout()"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Edit Metadata",
  "colab": {
   "collapsed_sections": [
    "A7p7eaeO1_bb",
    "ODt4TiKL1_cX",
    "Hmc3x1n-1_cr"
   ],
   "name": "py_double_ml_multiway_cluster.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "dml_base",
   "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": "3.10.8"
  },
  "vscode": {
   "interpreter": {
    "hash": "04413e1efd13b5e2aaedccc5facec5686292d28983ef24fa021a12c73bd6369e"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}