Feedback_analyse_from_DataStore.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Feedback analyse from DataStore.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/ZackAkil/rotting-zombie-model/blob/master/Feedback_analyse_from_DataStore.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "__dj6ULTiSrp",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Feedback analyse\n",
        "We will look at how you can extract data out of DataStore and check if the model is rotting."
      ]
    },
    {
      "metadata": {
        "id": "cPnQCcwziebG",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# install DataStore python library\n",
        "!pip install -qq google-cloud-datastore"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "SPH9J5l0ilAi",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from google.cloud import datastore"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "3n-dLTuniogz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# autheticate to our Google Cloud Platform account\n",
        "from google.colab import auth\n",
        "auth.authenticate_user()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "xkX3caYpjDod",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Connect to DataStore"
      ]
    },
    {
      "metadata": {
        "id": "Erf4Bcltitim",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "PROJECT_ID = 'YOUR PROJECT ID'\n",
        "\n",
        "# connect to DataStore\n",
        "client = datastore.Client(PROJECT_ID)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "j-1nFBetjICx",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Query DataStore"
      ]
    },
    {
      "metadata": {
        "id": "jItfTSdsi7o2",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# query for all prediction-feedback items\n",
        "query = client.query(kind='prediction-feedback')\n",
        "\n",
        "# order by the time field\n",
        "query.order = ['time']\n",
        "\n",
        "# fetch the items (returns an iterator so we will empty it into a list)\n",
        "data = list(query.fetch())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "UYHwm1khjZKn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "00ef2b00-a17a-46d7-fd9f-23e5bfff188f"
      },
      "cell_type": "code",
      "source": [
        "# lets look at a single item\n",
        "data[0]"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Entity('prediction-feedback', 5679790782676992) {'prediction': 1, 'input data': [-0.8300105114555543, 0.3990742221560673, 1.9084475908892906, 0.3804372006233603], 'was correct': True, 'model': 'v1', 'time': datetime.datetime(2018, 10, 22, 14, 21, 2, 199917, tzinfo=<UTC>)}>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "metadata": {
        "id": "a64BJ8zEjjg4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "d7c882f8-773d-4af8-82a4-86615cffab17"
      },
      "cell_type": "code",
      "source": [
        "# look at single items fields\n",
        "print(data[0]['was correct'])\n",
        "print(data[0]['model'])\n",
        "print(data[0]['time'])\n",
        "print(data[0]['input data'])"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "True\n",
            "v1\n",
            "2018-10-22 14:21:02.199917+00:00\n",
            "[-0.8300105114555543, 0.3990742221560673, 1.9084475908892906, 0.3804372006233603]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "3H0LLikFlFcz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## What percentage of feedback were  'was correct' was True"
      ]
    },
    {
      "metadata": {
        "id": "fXU1K_bkkshv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "35366a53-1add-484f-f699-31c689b36d8a"
      },
      "cell_type": "code",
      "source": [
        "number_of_items = len(data)\n",
        "number_of_was_correct = len([d for d in data if d['was correct']])\n",
        "\n",
        "\n",
        "print('{0} total number of items, {1} total number of items were \"was correct\" is set to True'.format(number_of_items, \n",
        "                                                                                                      number_of_was_correct))"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "250 total number of items, 210 total number of items were \"was correct\" is set to True\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "aOuBMXb-l0yg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ccf8ee21-7b07-45a8-8c2d-245c414af5a3"
      },
      "cell_type": "code",
      "source": [
        "# what ratio are correct i.e the accuracy of the model\n",
        "number_of_was_correct/number_of_items"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.84"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "metadata": {
        "id": "HUCKTbotmrLd",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Let's check to see if the accuracy is consistent over time"
      ]
    },
    {
      "metadata": {
        "id": "rS2sXd-_l9aV",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def plot_performance(data, window = 50):\n",
        "  '''\n",
        "  Will calculate the accuracy using a sliding window on the data.\n",
        "  '''\n",
        "  \n",
        "  perf = []\n",
        "\n",
        "  for i in range(len(data) - window):\n",
        "    data_slice = data[i:i+window]\n",
        "    accuracy = len([d for d in data_slice if d['was correct']]) / len(data_slice)\n",
        "    perf.append(accuracy)\n",
        "\n",
        "  plt.ylabel('Accuracy')\n",
        "  plt.plot(perf)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "dFiDSRm9rn7w",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 351
        },
        "outputId": "4538860a-5ad9-4100-dcdc-370765d1d556"
      },
      "cell_type": "code",
      "source": [
        "plot_performance(data)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFOCAYAAACfRYSnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkRElCL+L59wLo+efG/SAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f10c0d27dd8>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Ac91pP_Qsr24",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### So our model appears to have recently gotten a lot worse in performance  \n",
        "\n",
        "## Let's look at the corresponding input data that was submitted  "
      ]
    },
    {
      "metadata": {
        "id": "k-Zkx_6br2kM",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "def plot_input_data(data, start_index, window_size = 50):\n",
        "  \n",
        "  data_window = data[start_index : start_index + window_size]\n",
        "  \n",
        "  points = [d['input data'] for d in data_window]\n",
        "  correct = [int(d['was correct']) for d in data_window]\n",
        "\n",
        "  np_points = np.array(points)\n",
        "\n",
        "  plt.scatter(np_points[:,0], np_points[:,1], c=correct, cmap='bwr_r')\n",
        "  plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "bikoT4wFtUuq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "outputId": "3a18dbef-96e3-49cd-b977-8561e76648a2"
      },
      "cell_type": "code",
      "source": [
        "#@title Inspect input data { run: \"auto\" }\n",
        "start = 168 #@param {type:\"slider\", min:0, max:200, step:1}\n",
        "\n",
        "plot_input_data(data, start)"
      ],
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd8AAAFKCAYAAABcq1WoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f10c0c8a2b0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "KIjmSGOyu5yV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 707
        },
        "outputId": "a329538f-8073-42e5-edda-c2a766f4e6b7"
      },
      "cell_type": "code",
      "source": [
        "plt.title('High accuracy input data')\n",
        "plot_input_data(data, 0)\n",
        "\n",
        "plt.title('Low accuracy input data')\n",
        "plot_input_data(data, 170)"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd8AAAFZCAYAAADZ6SWdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f10c0d2d278>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd8AAAFZCAYAAADZ6SWdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f10c8895b00>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "N3i4rnKKtp9V",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "We can see that the input data was very different in the more recent feedback (were it got a low accuracy).\n",
        "\n",
        "We can guesse that the zombie population has mutated and we'll need to retrain our model on the new data collected (plus maybe go out and collect some more data of our own to confirm this hypothosis)."
      ]
    },
    {
      "metadata": {
        "id": "9_fD4d3rx-ak",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Let's collect and save the data from DataStore into a new data set"
      ]
    },
    {
      "metadata": {
        "id": "BhnbZBsQtWLx",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# performance dropped after the 125th item\n",
        "new_data = data[125:]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "iNQcZs8mxkRd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "db21bbb0-dda4-4a0b-a1ee-025772eb38c8"
      },
      "cell_type": "code",
      "source": [
        "# extract the input data\n",
        "new_x = np.array([d['input data'] for d in new_data])\n",
        "new_x.shape"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(125, 4)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "metadata": {
        "id": "9sdeyJ4HyWmg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "5c9ec7c9-becd-49cd-b410-dc76a82db0b2"
      },
      "cell_type": "code",
      "source": [
        "# extract the target y\n",
        "# BUT remember to flip it if the user said it was and incorrect prediction\n",
        "\n",
        "new_y = []\n",
        "\n",
        "for item in new_data:\n",
        "  \n",
        "  if item['was correct']:\n",
        "    # if the prediction was correct then just save the prediction as is    \n",
        "    new_y.append(item['prediction'])\n",
        "    \n",
        "  else:\n",
        "    # if it was incorrect then flip the prediction\n",
        "    if item['prediction'] == 0:\n",
        "      new_y.append(1)\n",
        "    else:\n",
        "      new_y.append(0)\n",
        "      \n",
        "new_y = np.array(new_y)\n",
        "\n",
        "new_y.shape"
      ],
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(125,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 102
        }
      ]
    },
    {
      "metadata": {
        "id": "9Zxj4yri0WTw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "37111a84-eae2-416e-88f8-08e1ec1ca637"
      },
      "cell_type": "code",
      "source": [
        "from sklearn.externals import joblib\n",
        "\n",
        "# save data\n",
        "joblib.dump({\"x\":new_x, \"y\":new_y}, \"zombie_data_2.joblib\" )"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['zombie_data_2.joblib']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 105
        }
      ]
    }
  ]
}