bindings/jupyter/notebooks/feature-list-view.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature List View\n",
    "\n",
    "## Usage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "92fae3525cd246e4bd6bdaabca12fd98",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "FeatureListView(props='{\"data\": [{\"name\": \"dummy-categorical-feature\", \"type\": \"categorical\", \"domain\": [\"Alab…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import sys, json, math\n",
    "from mlvis import FeatureListView\n",
    "from random import uniform, gauss\n",
    "from IPython.display import display\n",
    "if sys.version_info[0] < 3:\n",
    "    import urllib2 as url\n",
    "else:\n",
    "    import urllib.request as url\n",
    "    \n",
    "def generate_random_steps(k):\n",
    "    randoms = [uniform(0, 1) / 2 for i in range(0, k)]\n",
    "    steps =  [0] * (k - 1)\n",
    "    t = 0\n",
    "    for i in range(0, k - 1):\n",
    "        steps[i] = t + (1 - t) * randoms[i]\n",
    "        t = steps[i]\n",
    "    return steps + [1]\n",
    "\n",
    "def generate_categorical_feature(states):\n",
    "    size = len(states)\n",
    "    distro_a = [uniform(0, 1) for i in range(0, size)] \n",
    "    distro_b = [uniform(0, 1) for i in range(0, size)]\n",
    "    return {\n",
    "        'name': 'dummy-categorical-feature',\n",
    "        'type': 'categorical',\n",
    "        'domain': list(states.values()),\n",
    "        'distributions': [distro_a, distro_b],\n",
    "        'distributionNormalized': [distro_a, distro_b],\n",
    "        'colors': ['#47B274', '#6F5AA7'],\n",
    "        'divergence': uniform(0, 1)\n",
    "    }\n",
    "\n",
    "def generate_numerical_feature():\n",
    "    domain_size = 100\n",
    "    distro_a = [uniform(0, 1) for i in range(0, domain_size)]\n",
    "    distro_b = [uniform(0, 1) for i in range(0, domain_size)]\n",
    "    return {\n",
    "        'name': 'dummy-categorical-numerical',\n",
    "        'type': 'numerical',\n",
    "        'domain': generate_random_steps(domain_size),\n",
    "        'distributions': [distro_a, distro_b],\n",
    "        'distributionNormalized': [distro_a, distro_b],\n",
    "        'colors': ['#47B274', '#6F5AA7'],\n",
    "        'divergence': uniform(0, 1)\n",
    "    }\n",
    "\n",
    "def generate_random_categorical_values(states):\n",
    "    k = 10000\n",
    "    values = [None] * k\n",
    "    domain = list(states.values())\n",
    "    size = len(states)\n",
    "    for i in range(0, k):\n",
    "        d = int(math.floor(uniform(0, 1) * size))\n",
    "        values[i] = domain[d]\n",
    "    return values\n",
    "\n",
    "def generate_raw_categorical_feature(states):\n",
    "    return {\n",
    "        'name': 'dummy-raw-categorical-feature',\n",
    "        'type': 'categorical',\n",
    "        'values': [generate_random_categorical_values(states),\n",
    "                   generate_random_categorical_values(states)]\n",
    "    }\n",
    "\n",
    "def generate_raw_numerical_feature():\n",
    "    return {\n",
    "        'name': 'dummy-raw-numerical-feature',\n",
    "        'type': 'numerical',\n",
    "        'values': [\n",
    "            [gauss(2, 0.5) for i in range(0, 2500)],\n",
    "            [gauss(0, 1) for i in range(0, 7500)]\n",
    "        ]\n",
    "    }\n",
    "\n",
    "# load the US states data\n",
    "PREFIX = 'https://d1a3f4spazzrp4.cloudfront.net/mlvis/'\n",
    "response = url.urlopen(PREFIX + 'jupyter/states.json')\n",
    "states = json.loads(response.read().decode())\n",
    "\n",
    "# Randomly generate the data for the feature list view\n",
    "categorical_feature = generate_categorical_feature(states)\n",
    "raw_categorical_feature = generate_raw_categorical_feature(states)\n",
    "numerical_feature = generate_numerical_feature()\n",
    "raw_numerical_feature = generate_raw_numerical_feature()\n",
    "data = [categorical_feature, raw_categorical_feature, numerical_feature, raw_numerical_feature]\n",
    "\n",
    "feature_list_view = FeatureListView(props={\"data\": data, \"width\": 1000})\n",
    "\n",
    "display(feature_list_view) "
   ]
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}