39_Clustering_K_Means_Scikit.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 39 - Clustering: K_Means con Scikit\n",
    "\n",
    "\n",
    "* En este Notebook vamos a resolver un problema de ***Clusterización usando la librería de Scikit-Learn***.\n",
    "\n",
    "\n",
    "* Para resolver el problema de Clusterización vamos a utilizar la clase **\"KMeans()\"**, cuya documentación esta en el siguiente enlace:<br>\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html\n",
    "\n",
    "\n",
    "* El ejemplo con el que vamos a trabajar, consiste en ***agrupar una serie de flores*** de las que sabemos ***los tamaños de su sépalo***\n",
    "\n",
    "\n",
    "* El objetivo es ***encontrar 'K' Clusters que minimicen la Inercia*** (la suma de las distancias de cada elemento al centroide de su cluster):\n",
    "<span></span><br><br>\n",
    "<span style=\"font-size:20px\">$$\\underset{K}{argmin} \\sum_{j=1}^{k}\\sum_{x_i \\in  c_j}\\left \\| x_i - \\mu  \\right \\|^{2}$$</span>\n",
    "\n",
    "\n",
    "* Para resolver este problema vamos a realizar los siguientes pasos:\n",
    "<span></span><br>\n",
    "    1. [Carga de datos](#M1)\n",
    "<span></span><br>\n",
    "    2. [Cambio de estructura de datos a numpy](#M2)\n",
    "<span></span><br>\n",
    "    3. [Visualización de los datos](#M3)\n",
    "<span></span><br>\n",
    "    4. [Creacción del modelo (hipótesis) y ajuste](#M4)\n",
    "<span></span><br>\n",
    "    5. [Obtención del modelo (hipótesis)](#M5)\n",
    "<span></span><br>\n",
    "    6. [Visualización de los resultados](#M6)\n",
    "\n",
    "\n",
    "##### Importante: En los problemas de clusterización no conocemos el dominio de salida, por lo que para el problema que se plantea no vamos a saber el tipo de flor, solo vamos a saber los tamaños de las flores.\n",
    "\n",
    "\n",
    "* En primer lugar vamos a importar las librería necesarias:\n",
    "\n",
    "<hr>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.cluster import KMeans\n",
    "%matplotlib inline "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M1\">1. Carga de datos</a>\n",
    "\n",
    "\n",
    "* El primer paso que vamos a realizar es el de cargar los datos que los tenemos en el siguiente fichero:\n",
    "    - ***DataSet***: \"./data/iris/iris_para_clustering.csv\"\n",
    "\n",
    "\n",
    "* Para ello ***leeremos el csv con pandas*** (pasandolo a un dataframe), teniendo las siguientes columnas que nos interesan para el problema a resolver:\n",
    "    - ***sepal_length***: El \"largo\" del Sepalo\n",
    "    - ***sepal_width***: El \"ancho\" del Sepalo\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "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>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>4.6</td>\n",
       "      <td>3.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.0</td>\n",
       "      <td>3.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>6.4</td>\n",
       "      <td>2.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>6.3</td>\n",
       "      <td>2.9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     sepal_length  sepal_width\n",
       "6             4.6          3.4\n",
       "60            5.0          2.0\n",
       "4             5.0          3.6\n",
       "132           6.4          2.8\n",
       "103           6.3          2.9"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"./data/iris/iris_para_clustering.csv\")\n",
    "df[['sepal_length', 'sepal_width']].sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M2\">2. Cambio de estructura de datos a numpy\n",
    "    \n",
    "* Scikit necesita como entrada para el ajuste (o entrenamiento) de los modelos de Clustering los datos del Dataset de la siguiente forma:\n",
    "<span></span><br><br>\n",
    "    - ***Variables de entrada X***: ***Array de Arrays de numpy*** en el que cada array tiene las variables de entrada de un elemento.\n",
    "\n",
    "\n",
    "* Por tanto, vamos a pasar los datos las flores a una nueva estructura de datos para que scikit-learn puedan trabajar con los datos:\n",
    "<span></span><br><br>\n",
    "    - Las ***columnas sepal_length y sepal_width*** representan las variables de entrada y las vamos a pasar a un array de array (***numpy array***)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df[['sepal_length', 'sepal_width']].values   # Obtengo el alto y ancho del sepalo"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M3\">3. Visualización de los datos\n",
    "    \n",
    "* Vamos a pasar a ver la relación que hay entre el alto y ancho del sepalo de las flores, para ver si podemos apreciar de manera visual los grupo de flores que puede haber.\n",
    "\n",
    "\n",
    "##### NOTA: Este es un ejercicio con finalidades didácticas y se pueden hacer estos ejercicios introspectivos. En los casos reales el clustering se hace sobre muchas más variables y no se pueden apreciar de esta manera las relaciones entre variables.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fbcbf152908>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.scatterplot(x=\"sepal_length\", y=\"sepal_width\", data = df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M4\">4. Creacción del modelo (hipótesis) y ajuste\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KMeans(init='random', max_iter=50, n_clusters=3)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creamos un objeto de la clase \"KMeans()\"\n",
    "kmeans = KMeans(n_clusters=3, max_iter=50, init=\"random\")\n",
    "\n",
    "# Ajustamos el modelo\n",
    "kmeans.fit(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M5\">5. Obtención del modelo (hipótesis)\n",
    "\n",
    "\n",
    "* Tras ajustar el modelo podemos obtener los siguientes valores:\n",
    "\n",
    "    1. ***El valor de cada uno de los centroides***: Esto lo hacemos llamando al atributo ***\"cluster_centers_\"*** de la clase Kmeans.\n",
    "    2. ***El valor de la inercia***: Esto lo hacemos llamando al atributo ***\"inertia_\"*** de la clase Kmeans.\n",
    "    3. ***A que cluster pertenece cada item***:  Esto lo hacemos llamando al atributo ***\"labels_\"*** de la clase Kmeans."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster 1\n",
      "\tCentroide: [5.006 3.418]\n",
      "Cluster 2\n",
      "\tCentroide: [6.81276596 3.07446809]\n",
      "Cluster 3\n",
      "\tCentroide: [5.77358491 2.69245283]\n"
     ]
    }
   ],
   "source": [
    "# Valores de los centroides\n",
    "centroides = kmeans.cluster_centers_\n",
    "for index, centroide in enumerate(centroides):\n",
    "    print(\"Cluster {}\".format(index + 1))\n",
    "    print(\"\\tCentroide: {}\".format(centroide))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "37.1237021276596"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Valor de la inercia\n",
    "kmeans.inertia_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 1, 1, 1, 2, 1, 2, 1, 2, 1, 2, 2, 2, 2, 2, 2, 1,\n",
       "       2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2,\n",
       "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 1, 1, 1, 1, 2, 1, 1, 1,\n",
       "       1, 1, 1, 2, 2, 1, 1, 1, 1, 2, 1, 2, 1, 2, 1, 1, 2, 2, 1, 1, 1, 1,\n",
       "       1, 2, 2, 1, 1, 1, 2, 1, 1, 1, 2, 1, 1, 1, 2, 1, 1, 2], dtype=int32)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Etiqueta del cluster a la que se asigna cada elemento\n",
    "kmeans.labels_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "\n",
    "## <a name=\"M6\">6. Visualización de los resultados\n",
    "    \n",
    "\n",
    "* A continuación pasamos a representar en un espacio de dos dimensiones cómo agruparía nuestro modelo en ese espacio 2D y cómo asigna a cada Cluster los elementos del Dataset con el que ha entrenado:\n",
    "\n",
    "\n",
    "* En primer lugar vamos a definir una malla de puntos (\"meshgrid\") que se utilizará para pintar las regiones de cada cluster. El objetivo de construir esta malla es la de predecir cada punto de la malla como si fuesen los datos de entrada de una flor con esos valores de alto y ancho.\n",
    "\n",
    "\n",
    "* Para ello miramos los valores máximos y mínimos de los tamaños de la flor para acotar la región a pintar y posteriormente generamos el conjunto de puntos a predecir (asignar en el Cluster), llamando a la función ***\"predict(X)\"***"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5\n",
    "y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5\n",
    "h = .01\n",
    "xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
    "\n",
    "# Predicción de cada uno de los puntos de la malla\n",
    "Z = kmeans.predict(np.c_[xx.ravel(), yy.ravel()])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Pasamos a pintar las regiones de cada Cluster así como los datos del dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Z = Z.reshape(xx.shape)\n",
    "plt.figure(1, figsize=(8, 6))\n",
    "\n",
    "# Dibujamos la malla con los diferentes colores\n",
    "plt.pcolormesh(xx, yy, Z, cmap=plt.cm.Pastel1)\n",
    "\n",
    "\n",
    "# Dibujamos los puntos (flores) del dataset\n",
    "labels = kmeans.labels_\n",
    "plt.scatter(X[:, 0], X[:, 1], c=labels, edgecolors='k', cmap=plt.cm.viridis)\n",
    "\n",
    "# Dibujamos los Centroides\n",
    "centroides = kmeans.cluster_centers_\n",
    "plt.plot(centroides[:, 0], centroides[:, 1], 'o', markeredgecolor='k', markersize=14)\n",
    "\n",
    "plt.title('Clustering')\n",
    "plt.xlabel('Sepal length')\n",
    "plt.ylabel('Sepal width')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "\n",
    "*Este Notebook ha sido desarrollado por **Ricardo Moya García** y registrado en Safe Creative como ***Atribución-NoComercial-CompartirIgual***.*\n",
    "\n",
    "<img src=\"./imgs/CC_BY-NC-SA.png\" alt=\"CC BY-NC\">"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}