pydlm_hello.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "pydlm_hello.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyO3V9manzl3cpna8KicHH56",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/microprediction/timeseries-notebooks/blob/main/pydlm_hello.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pxPf4jy3Omj_"
      },
      "source": [
        "!pip install pydlm"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-KgmO47lOr3S"
      },
      "source": [
        "!pip install microprediction"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p2bh6KxE88So"
      },
      "source": [
        "# Hello world example\n",
        "See https://www.microprediction.com/blog/popular-timeseries-packages for more packages"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DUhhIeHhOutZ"
      },
      "source": [
        "from microprediction import MicroReader\n",
        "mr = MicroReader()\n",
        "YS = mr.get_lagged_values(name='emojitracker-twitter-face_with_medical_mask.json')[:100]"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ODDpGjVDPDwS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "10974d08-a477-4a6f-b2f0-2d066ba10f25"
      },
      "source": [
        "from pydlm import dlm, trend, seasonality\n",
        "def run(ys:[float])->[float]:\n",
        "     \"\"\" Compute list of priors for observations ys \"\"\"\n",
        "     model = dlm([],printInfo=False) + trend(1) + seasonality(24)\n",
        "     xs = [0] # First guess doesn't matter\n",
        "     for t,y in enumerate(ys[:-1]):\n",
        "        model.append([y])\n",
        "        model.fitForwardFilter()\n",
        "        x_mean,x_var = model.predict()\n",
        "        xs.append(x_mean[0,0])\n",
        "     return xs\n",
        "XS = run(YS)\n",
        "len(XS)\n"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "100"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "UsQalSXqS6a4",
        "outputId": "f35b4b55-d509-4951-b981-c7115cff4f71"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(YS[75:85],'*b')\n",
        "plt.plot(XS[75:85],'g')\n",
        "plt.legend(['data','prior'])"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f2c29422470>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}