data-x-crash-course.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![data-x](http://oi64.tinypic.com/o858n4.jpg)\n",
    "\n",
    "---\n",
    "# Data-X Crash Course\n",
    "\n",
    "\n",
    "**Author:** Alexander Fred-Ojala\n",
    "\n",
    "\n",
    "**License Agreement:** Feel free to do whatever you want with this code\n",
    "\n",
    "___"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Outline\n",
    "\n",
    "> ### 0. Note on tools\n",
    "> ### 1. Intro to Jupyter Notebooks\n",
    "> ### 2. Numpy, pandas and matplotlib\n",
    "> ### 3. Scikit-learn and Machine Learning\n",
    "> ### 4. Tensorflow / Keras for Deep Learning\n",
    "\n",
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 0. Note on tools\n",
    "\n",
    "### Almost every problem can be framed as a data problem\n",
    "\n",
    "This is a real world example of how you'd solve a Machine Learning prediciton problem.\n",
    "\n",
    "**Use Cases:**\n",
    "- Discover churn risk of customers\n",
    "- Predict optimal price levels (investments / retail)\n",
    "- Predict future revenues\n",
    "- Build recommendation systems\n",
    "- Customer value scoring\n",
    "- Fraud detection\n",
    "- Customer insights (characteristics)\n",
    "- Predict sentiment of text\n",
    "- Object detecton in images\n",
    "- etc etc..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Why Python?\n",
    "\n",
    "Python is general purpose and can do Software development, Web development, AI. Python has experienced incredible growth over the last couple of years, and many of the state of the art Machine Learning libraries being developed today have support for Python.\n",
    "\n",
    "<img src='https://zgab33vy595fw5zq-zippykid.netdna-ssl.com/wp-content/uploads/2017/09/growth_major_languages-1-1400x1200.png' width=400px></img>\n",
    "\n",
    "Source: https://stackoverflow.blog/2017/09/06/incredible-growth-python/\n",
    "\n",
    "# Everything is free!\n",
    "\n",
    "The best software today is open source and it's also enterprise-ready. Anyone can download and use them for free (even for business purposes).\n",
    "\n",
    "**Examples of great, free AI libraries:**\n",
    "* Anaconda\n",
    "* Google's TensorFlow\n",
    "* Scikit-learn\n",
    "* Pandas\n",
    "* Keras\n",
    "* Matplotlib\n",
    "* SQL\n",
    "* Spark\n",
    "* Numpy\n",
    "\n",
    "## State-of-the-Art algorithms\n",
    "\n",
    "No matter what algorithm you want to use (Linear Regression, Random Forests, Neural Networks, or Deep Learning), **all of the latest methods are implemented optimized for Python**.\n",
    "\n",
    "## Big Data\n",
    "\n",
    "Python code can run on any computer. Therefore, you can scale your computations and utilize for example cloud resources to run big data jobs.\n",
    "\n",
    "**Great tools for Big Data:**\n",
    "- Spark\n",
    "- Databricks\n",
    "- Hadoop / MapReduce\n",
    "- Kafka\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Jupyter Notebooks\n",
    "\n",
    "1. Markdown cells to annotate work\n",
    "2. Code cells to run code"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "### 1.1 Markdown cells\n",
    "\n",
    "Format text **bold**, _italic_, `code`, \n",
    "\n",
    "#### Latex:\n",
    "$$\\int^{\\inf}_{-\\inf}\\text{erf}^{2\\pi x}\\text{dx} = \\frac{1}{\\sqrt{2\\pi}}$$\n",
    "\n",
    "#### Images:\n",
    "<img src='imgs/dx_logo.png' width='300px'>\n",
    "\n",
    "#### Bullet lists:\n",
    "- Unordered\n",
    "- list\n",
    "\n",
    "1. Ordered\n",
    "2. list\n",
    "\n",
    "<br><br>\n",
    "\n",
    "**Find a lot of useful Markdown commands here:** https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.2 Code cells\n",
    "\n",
    "Run Python code sequentially"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1+2\n",
    "3+2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def add(x,y):\n",
    "    return(x+y)\n",
    "\n",
    "add(2,2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " best UC Berkeley class!! Data-X is the\r"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m--------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m<ipython-input-8-aaf3d6ac814c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mend\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'\\r'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mflush\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m         \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msleep\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m.1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "import time\n",
    "s = 'Data-X is the best UC Berkeley class!! '\n",
    "l = len(s)\n",
    "\n",
    "while True:\n",
    "    for i in range(l):\n",
    "        print(s[-i:]+s[:-i],end='\\r',flush=True)\n",
    "        time.sleep(.1)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### For more info about Jupyter Notebooks and Python see: [The Intro Notebook](../01-intro/python-jupyter-basics_short_afo.ipynb)\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Python's data science trinity:\n",
    "### Numpy & Pandas (& Matplotlib)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.1 Numpy\n",
    "Python's Numerical computation library. Great for multidim arrays, data manipulation, generation, and broadcasting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1, 2, 3],\n",
       "       [4, 5, 6]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array([[1,2,3],[4,5,6]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.74562867,  0.0486068 ],\n",
       "       [-0.67487283, -0.25029024],\n",
       "       [ 1.15626201, -0.57527435],\n",
       "       [-0.97639403,  0.71276788],\n",
       "       [-1.18506434,  0.13723881]])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.randn(10).reshape(5,2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1., 0., 0., 0., 0.],\n",
       "       [0., 1., 0., 0., 0.],\n",
       "       [0., 0., 1., 0., 0.],\n",
       "       [0., 0., 0., 1., 0.],\n",
       "       [0., 0., 0., 0., 1.]])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.eye(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.        ,  1.57079633,  3.14159265,  4.71238898,  6.28318531,\n",
       "        7.85398163,  9.42477796, 10.99557429, 12.56637061])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = np.linspace(0,4*np.pi,9)\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.,  1.,  0., -1., -0.,  1.,  0., -1., -0.])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.round(np.sin(x),10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = np.linspace(0,10,100)\n",
    "y = np.sin(x) + np.random.normal(0,.1,100)\n",
    "plt.plot(x,y);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### For more info about Jupyter Notebooks and Python see: [The Numpy Notebook](../02-AI-stack/numpy.ipynb)\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2 Pandas\n",
    "Python's tabular data library (my favorite!)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.180392</td>\n",
       "      <td>-1.323788</td>\n",
       "      <td>-0.709621</td>\n",
       "      <td>-0.076805</td>\n",
       "      <td>-1.108328</td>\n",
       "      <td>-1.249897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.779359</td>\n",
       "      <td>2.064447</td>\n",
       "      <td>-0.783838</td>\n",
       "      <td>-0.524770</td>\n",
       "      <td>-0.004255</td>\n",
       "      <td>1.128987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.489815</td>\n",
       "      <td>-0.305573</td>\n",
       "      <td>0.196747</td>\n",
       "      <td>-0.524721</td>\n",
       "      <td>0.476839</td>\n",
       "      <td>-1.897621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.080633</td>\n",
       "      <td>-1.045533</td>\n",
       "      <td>0.060527</td>\n",
       "      <td>1.811168</td>\n",
       "      <td>-1.012252</td>\n",
       "      <td>-2.698729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.894994</td>\n",
       "      <td>-1.684144</td>\n",
       "      <td>-0.369236</td>\n",
       "      <td>-0.387430</td>\n",
       "      <td>-1.248294</td>\n",
       "      <td>0.454361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.952431</td>\n",
       "      <td>-0.699643</td>\n",
       "      <td>-0.614335</td>\n",
       "      <td>-0.981936</td>\n",
       "      <td>-0.051740</td>\n",
       "      <td>1.632168</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>-0.194322</td>\n",
       "      <td>-2.021760</td>\n",
       "      <td>1.169097</td>\n",
       "      <td>-0.819832</td>\n",
       "      <td>-0.313248</td>\n",
       "      <td>-0.978625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.413214</td>\n",
       "      <td>-0.689088</td>\n",
       "      <td>0.867004</td>\n",
       "      <td>0.535125</td>\n",
       "      <td>0.069258</td>\n",
       "      <td>-1.606911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>-0.933705</td>\n",
       "      <td>-0.708602</td>\n",
       "      <td>0.896151</td>\n",
       "      <td>1.439796</td>\n",
       "      <td>-0.436123</td>\n",
       "      <td>0.906631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>-0.326036</td>\n",
       "      <td>0.080049</td>\n",
       "      <td>-1.281727</td>\n",
       "      <td>-0.312406</td>\n",
       "      <td>0.860485</td>\n",
       "      <td>-0.190859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>-0.037969</td>\n",
       "      <td>0.613154</td>\n",
       "      <td>-0.780651</td>\n",
       "      <td>1.582745</td>\n",
       "      <td>2.679898</td>\n",
       "      <td>-0.794721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>-0.876523</td>\n",
       "      <td>-1.344248</td>\n",
       "      <td>-1.814433</td>\n",
       "      <td>0.169623</td>\n",
       "      <td>0.745099</td>\n",
       "      <td>-0.751244</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.323897</td>\n",
       "      <td>-0.076347</td>\n",
       "      <td>0.405778</td>\n",
       "      <td>-0.789652</td>\n",
       "      <td>0.321080</td>\n",
       "      <td>-0.641060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>-1.038246</td>\n",
       "      <td>-0.096380</td>\n",
       "      <td>-0.048636</td>\n",
       "      <td>-1.122800</td>\n",
       "      <td>0.119283</td>\n",
       "      <td>1.005498</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.767552</td>\n",
       "      <td>-0.885713</td>\n",
       "      <td>-0.550673</td>\n",
       "      <td>0.815845</td>\n",
       "      <td>-0.256274</td>\n",
       "      <td>-0.546862</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1.472331</td>\n",
       "      <td>0.904508</td>\n",
       "      <td>0.243524</td>\n",
       "      <td>-1.692492</td>\n",
       "      <td>-0.190622</td>\n",
       "      <td>1.280605</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.477851</td>\n",
       "      <td>-0.613560</td>\n",
       "      <td>-2.026870</td>\n",
       "      <td>-0.019286</td>\n",
       "      <td>-0.018444</td>\n",
       "      <td>0.767479</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>-0.986643</td>\n",
       "      <td>0.282039</td>\n",
       "      <td>-0.373597</td>\n",
       "      <td>0.042560</td>\n",
       "      <td>1.589031</td>\n",
       "      <td>0.988990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>-0.862332</td>\n",
       "      <td>-0.741035</td>\n",
       "      <td>0.159147</td>\n",
       "      <td>-0.877932</td>\n",
       "      <td>-0.129239</td>\n",
       "      <td>0.525744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>-0.022209</td>\n",
       "      <td>-0.641405</td>\n",
       "      <td>-0.515822</td>\n",
       "      <td>-0.411878</td>\n",
       "      <td>-1.150848</td>\n",
       "      <td>1.951283</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0         1         2         3         4         5\n",
       "0   0.180392 -1.323788 -0.709621 -0.076805 -1.108328 -1.249897\n",
       "1   1.779359  2.064447 -0.783838 -0.524770 -0.004255  1.128987\n",
       "2   0.489815 -0.305573  0.196747 -0.524721  0.476839 -1.897621\n",
       "3   0.080633 -1.045533  0.060527  1.811168 -1.012252 -2.698729\n",
       "4  -1.894994 -1.684144 -0.369236 -0.387430 -1.248294  0.454361\n",
       "5   0.952431 -0.699643 -0.614335 -0.981936 -0.051740  1.632168\n",
       "6  -0.194322 -2.021760  1.169097 -0.819832 -0.313248 -0.978625\n",
       "7   1.413214 -0.689088  0.867004  0.535125  0.069258 -1.606911\n",
       "8  -0.933705 -0.708602  0.896151  1.439796 -0.436123  0.906631\n",
       "9  -0.326036  0.080049 -1.281727 -0.312406  0.860485 -0.190859\n",
       "10 -0.037969  0.613154 -0.780651  1.582745  2.679898 -0.794721\n",
       "11 -0.876523 -1.344248 -1.814433  0.169623  0.745099 -0.751244\n",
       "12  0.323897 -0.076347  0.405778 -0.789652  0.321080 -0.641060\n",
       "13 -1.038246 -0.096380 -0.048636 -1.122800  0.119283  1.005498\n",
       "14  0.767552 -0.885713 -0.550673  0.815845 -0.256274 -0.546862\n",
       "15  1.472331  0.904508  0.243524 -1.692492 -0.190622  1.280605\n",
       "16  0.477851 -0.613560 -2.026870 -0.019286 -0.018444  0.767479\n",
       "17 -0.986643  0.282039 -0.373597  0.042560  1.589031  0.988990\n",
       "18 -0.862332 -0.741035  0.159147 -0.877932 -0.129239  0.525744\n",
       "19 -0.022209 -0.641405 -0.515822 -0.411878 -1.150848  1.951283"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame(np.random.randn(120).reshape(20,6))\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfg = pd.read_csv('./02-AI-stack/data/googl.csv').drop('Unnamed: 0',axis=1) # Google stock data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>Date</th>\n",
       "      <th>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>19-Jan-18</td>\n",
       "      <td>1138.03</td>\n",
       "      <td>1143.78</td>\n",
       "      <td>1132.50</td>\n",
       "      <td>1143.50</td>\n",
       "      <td>1527554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18-Jan-18</td>\n",
       "      <td>1139.35</td>\n",
       "      <td>1140.58</td>\n",
       "      <td>1124.46</td>\n",
       "      <td>1135.97</td>\n",
       "      <td>1374873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>17-Jan-18</td>\n",
       "      <td>1136.36</td>\n",
       "      <td>1139.32</td>\n",
       "      <td>1123.49</td>\n",
       "      <td>1139.10</td>\n",
       "      <td>1391510</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>16-Jan-18</td>\n",
       "      <td>1140.31</td>\n",
       "      <td>1148.88</td>\n",
       "      <td>1126.66</td>\n",
       "      <td>1130.70</td>\n",
       "      <td>1823100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12-Jan-18</td>\n",
       "      <td>1110.10</td>\n",
       "      <td>1131.30</td>\n",
       "      <td>1108.01</td>\n",
       "      <td>1130.65</td>\n",
       "      <td>1929306</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date     Open     High      Low    Close   Volume\n",
       "0  19-Jan-18  1138.03  1143.78  1132.50  1143.50  1527554\n",
       "1  18-Jan-18  1139.35  1140.58  1124.46  1135.97  1374873\n",
       "2  17-Jan-18  1136.36  1139.32  1123.49  1139.10  1391510\n",
       "3  16-Jan-18  1140.31  1148.88  1126.66  1130.70  1823100\n",
       "4  12-Jan-18  1110.10  1131.30  1108.01  1130.65  1929306"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfg.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfg = dfg.set_index(pd.to_datetime(dfg['Date'])).drop('Date',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-01-19</th>\n",
       "      <td>1138.03</td>\n",
       "      <td>1143.78</td>\n",
       "      <td>1132.50</td>\n",
       "      <td>1143.50</td>\n",
       "      <td>1527554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-18</th>\n",
       "      <td>1139.35</td>\n",
       "      <td>1140.58</td>\n",
       "      <td>1124.46</td>\n",
       "      <td>1135.97</td>\n",
       "      <td>1374873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-17</th>\n",
       "      <td>1136.36</td>\n",
       "      <td>1139.32</td>\n",
       "      <td>1123.49</td>\n",
       "      <td>1139.10</td>\n",
       "      <td>1391510</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-16</th>\n",
       "      <td>1140.31</td>\n",
       "      <td>1148.88</td>\n",
       "      <td>1126.66</td>\n",
       "      <td>1130.70</td>\n",
       "      <td>1823100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-12</th>\n",
       "      <td>1110.10</td>\n",
       "      <td>1131.30</td>\n",
       "      <td>1108.01</td>\n",
       "      <td>1130.65</td>\n",
       "      <td>1929306</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Open     High      Low    Close   Volume\n",
       "Date                                                   \n",
       "2018-01-19  1138.03  1143.78  1132.50  1143.50  1527554\n",
       "2018-01-18  1139.35  1140.58  1124.46  1135.97  1374873\n",
       "2018-01-17  1136.36  1139.32  1123.49  1139.10  1391510\n",
       "2018-01-16  1140.31  1148.88  1126.66  1130.70  1823100\n",
       "2018-01-12  1110.10  1131.30  1108.01  1130.65  1929306"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfg.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2017-04-28</th>\n",
       "      <td>929.00</td>\n",
       "      <td>935.90</td>\n",
       "      <td>923.22</td>\n",
       "      <td>924.52</td>\n",
       "      <td>3845857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-27</th>\n",
       "      <td>890.00</td>\n",
       "      <td>893.38</td>\n",
       "      <td>887.18</td>\n",
       "      <td>891.44</td>\n",
       "      <td>2342303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-26</th>\n",
       "      <td>891.39</td>\n",
       "      <td>892.99</td>\n",
       "      <td>885.15</td>\n",
       "      <td>889.14</td>\n",
       "      <td>1323297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-25</th>\n",
       "      <td>882.26</td>\n",
       "      <td>892.25</td>\n",
       "      <td>879.28</td>\n",
       "      <td>888.84</td>\n",
       "      <td>2038032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-24</th>\n",
       "      <td>868.44</td>\n",
       "      <td>879.96</td>\n",
       "      <td>866.11</td>\n",
       "      <td>878.93</td>\n",
       "      <td>1696483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-21</th>\n",
       "      <td>860.62</td>\n",
       "      <td>862.44</td>\n",
       "      <td>857.73</td>\n",
       "      <td>858.95</td>\n",
       "      <td>1172870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-20</th>\n",
       "      <td>859.74</td>\n",
       "      <td>863.93</td>\n",
       "      <td>857.50</td>\n",
       "      <td>860.08</td>\n",
       "      <td>1188843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-19</th>\n",
       "      <td>857.38</td>\n",
       "      <td>860.20</td>\n",
       "      <td>853.53</td>\n",
       "      <td>856.51</td>\n",
       "      <td>1080336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-18</th>\n",
       "      <td>852.54</td>\n",
       "      <td>857.39</td>\n",
       "      <td>851.25</td>\n",
       "      <td>853.99</td>\n",
       "      <td>936201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-17</th>\n",
       "      <td>841.38</td>\n",
       "      <td>855.64</td>\n",
       "      <td>841.03</td>\n",
       "      <td>855.13</td>\n",
       "      <td>1049125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-13</th>\n",
       "      <td>841.04</td>\n",
       "      <td>843.73</td>\n",
       "      <td>837.85</td>\n",
       "      <td>840.18</td>\n",
       "      <td>1073650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-12</th>\n",
       "      <td>838.46</td>\n",
       "      <td>843.72</td>\n",
       "      <td>837.59</td>\n",
       "      <td>841.46</td>\n",
       "      <td>1135775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-11</th>\n",
       "      <td>841.70</td>\n",
       "      <td>844.63</td>\n",
       "      <td>834.60</td>\n",
       "      <td>839.88</td>\n",
       "      <td>974343</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-10</th>\n",
       "      <td>841.54</td>\n",
       "      <td>846.74</td>\n",
       "      <td>840.79</td>\n",
       "      <td>841.70</td>\n",
       "      <td>1046215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-07</th>\n",
       "      <td>845.00</td>\n",
       "      <td>845.88</td>\n",
       "      <td>837.30</td>\n",
       "      <td>842.10</td>\n",
       "      <td>1111610</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-06</th>\n",
       "      <td>849.50</td>\n",
       "      <td>853.59</td>\n",
       "      <td>844.00</td>\n",
       "      <td>845.10</td>\n",
       "      <td>1533564</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-05</th>\n",
       "      <td>854.71</td>\n",
       "      <td>860.59</td>\n",
       "      <td>847.52</td>\n",
       "      <td>848.91</td>\n",
       "      <td>1855153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-04</th>\n",
       "      <td>848.00</td>\n",
       "      <td>853.00</td>\n",
       "      <td>847.50</td>\n",
       "      <td>852.57</td>\n",
       "      <td>1348494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-04-03</th>\n",
       "      <td>848.75</td>\n",
       "      <td>859.00</td>\n",
       "      <td>847.53</td>\n",
       "      <td>856.75</td>\n",
       "      <td>1969402</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Open    High     Low   Close   Volume\n",
       "Date                                               \n",
       "2017-04-28  929.00  935.90  923.22  924.52  3845857\n",
       "2017-04-27  890.00  893.38  887.18  891.44  2342303\n",
       "2017-04-26  891.39  892.99  885.15  889.14  1323297\n",
       "2017-04-25  882.26  892.25  879.28  888.84  2038032\n",
       "2017-04-24  868.44  879.96  866.11  878.93  1696483\n",
       "2017-04-21  860.62  862.44  857.73  858.95  1172870\n",
       "2017-04-20  859.74  863.93  857.50  860.08  1188843\n",
       "2017-04-19  857.38  860.20  853.53  856.51  1080336\n",
       "2017-04-18  852.54  857.39  851.25  853.99   936201\n",
       "2017-04-17  841.38  855.64  841.03  855.13  1049125\n",
       "2017-04-13  841.04  843.73  837.85  840.18  1073650\n",
       "2017-04-12  838.46  843.72  837.59  841.46  1135775\n",
       "2017-04-11  841.70  844.63  834.60  839.88   974343\n",
       "2017-04-10  841.54  846.74  840.79  841.70  1046215\n",
       "2017-04-07  845.00  845.88  837.30  842.10  1111610\n",
       "2017-04-06  849.50  853.59  844.00  845.10  1533564\n",
       "2017-04-05  854.71  860.59  847.52  848.91  1855153\n",
       "2017-04-04  848.00  853.00  847.50  852.57  1348494\n",
       "2017-04-03  848.75  859.00  847.53  856.75  1969402"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfg['2017-04']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>251.000000</td>\n",
       "      <td>251.000000</td>\n",
       "      <td>251.000000</td>\n",
       "      <td>251.000000</td>\n",
       "      <td>2.510000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>954.393944</td>\n",
       "      <td>960.413944</td>\n",
       "      <td>948.097490</td>\n",
       "      <td>954.886295</td>\n",
       "      <td>1.618960e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>79.500261</td>\n",
       "      <td>80.542821</td>\n",
       "      <td>78.730193</td>\n",
       "      <td>79.808157</td>\n",
       "      <td>6.907057e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>815.000000</td>\n",
       "      <td>822.390000</td>\n",
       "      <td>812.050000</td>\n",
       "      <td>815.240000</td>\n",
       "      <td>7.268980e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>872.105000</td>\n",
       "      <td>877.190000</td>\n",
       "      <td>867.945000</td>\n",
       "      <td>875.650000</td>\n",
       "      <td>1.191540e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>951.350000</td>\n",
       "      <td>959.560000</td>\n",
       "      <td>945.310000</td>\n",
       "      <td>954.720000</td>\n",
       "      <td>1.440982e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1005.335000</td>\n",
       "      <td>1008.630000</td>\n",
       "      <td>999.200000</td>\n",
       "      <td>1005.360000</td>\n",
       "      <td>1.787760e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1140.310000</td>\n",
       "      <td>1148.880000</td>\n",
       "      <td>1132.500000</td>\n",
       "      <td>1143.500000</td>\n",
       "      <td>5.933602e+06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Open         High          Low        Close        Volume\n",
       "count   251.000000   251.000000   251.000000   251.000000  2.510000e+02\n",
       "mean    954.393944   960.413944   948.097490   954.886295  1.618960e+06\n",
       "std      79.500261    80.542821    78.730193    79.808157  6.907057e+05\n",
       "min     815.000000   822.390000   812.050000   815.240000  7.268980e+05\n",
       "25%     872.105000   877.190000   867.945000   875.650000  1.191540e+06\n",
       "50%     951.350000   959.560000   945.310000   954.720000  1.440982e+06\n",
       "75%    1005.335000  1008.630000   999.200000  1005.360000  1.787760e+06\n",
       "max    1140.310000  1148.880000  1132.500000  1143.500000  5.933602e+06"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfg.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7.480526315789473"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(dfg['2017-02']['High']-dfg['2017-02']['Low'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfg['2017-02'][['Low','High']].plot()\n",
    "plt.title('Feb 2017 Low High for Google')\n",
    "plt.xlabel('Date')\n",
    "plt.ylabel('Price');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### For more info about Pandas see: [The Pandas Notebook](../02-AI-stack/pandas.ipynb)\n",
    "\n",
    "##### For more info about Pandas see: [The Matplotlib Notebook](../02-AI-stack/matplotlib.ipynb)\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Machine Learning with Scikit-Learn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = datasets.load_iris()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(data.DESCR[:1050])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.DataFrame(data.data,columns=data.feature_names)\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y = pd.Series(data.target,index=np.repeat(data.target_names,50))\n",
    "Y.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_val, Y_train, Y_val = train_test_split(X,Y,test_size=0.4,shuffle=True,random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train.shape, Y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val.shape, Y_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import LinearSVC"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression (Softmax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = LogisticRegression() # instantiate\n",
    "model.fit(X_train,Y_train) # train\n",
    "print('Accuracy on validation:',np.round(model.score(X_val,Y_val)*100,2),'%') # predict and evaluate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = RandomForestClassifier() # instantiate\n",
    "model.fit(X_train,Y_train) # train\n",
    "print('Accuracy on validation:',np.round(model.score(X_val,Y_val)*100,2),'%') # predict and evaluate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = LinearSVC() # instantiate\n",
    "model.fit(X_train,Y_train) #train\n",
    "print('Accuracy on validation:',np.round(model.score(X_val,Y_val)*100,2),'%') # predict and evaluate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### For more info about Scikit-learn & ML workflow see: [The Titanic Notebook](../04-tools-prediction-titanic/titanic.ipynb)\n",
    "\n",
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neural Networks / Deep Learning: Keras (TensorFlow backend)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(100,input_shape=[4],activation='relu'))\n",
    "model.add(Dense(50,activation='relu'))\n",
    "model.add(Dense(3,activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train = pd.get_dummies(Y_train)\n",
    "Y_val = pd.get_dummies(Y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.fit(X_train,Y_train,batch_size=10,epochs=100,verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "_,acc = model.evaluate(X_val,Y_val,verbose=0)\n",
    "print('Accuracy',np.round(acc*100),'%')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### For more info about Keras / TensorFlow see: [TensorFlow](../05-deep-learning/) and [Keras](../05-deep-learning/catsVSdogs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}