{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "%matplotlib inline"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "\nBinary Forecast Verification\n============================\n\nThis example demonstrates how to perform binary forecast verification.\n\nIf you are not familiar with the terms, you may wish to refer to the following\nresources:\n    \n* `Verification of categorical predictands by Anna Ghelli of ECMWF <https://www.swpc.noaa.gov/sites/default/files/images/u30/Verification%20of%20Categorical%20Predictands.pdf>`_\n* `Verification Measures by Weather Forecasting ... On-Line <https://www.wxonline.info/topics/verif2.html>`_\n\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Definitions\n--------------------------------------------------------\n\n\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Import all required modules and methods:\n\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false
      },
      "outputs": [],
      "source": [
        "# Python package to manage warning message\nimport warnings\n# Python package for time calculations\nimport pandas as pd\n# Python package for numerical calculations\nimport numpy as np\n# Python package for xarrays to read and handle netcdf data\nimport xarray as xr\n# swirlspy traditional binary verification package\nfrom swirlspy.ver.crosstab import contingency\n# swirlspy verification metric package\nimport swirlspy.ver.metric as mt\n\nwarnings.filterwarnings(\"ignore\")\n\n# Generating sample forecast and observed data\n# Dimensions: x, y, time (3 x 3 x 4)\nforecast_data = [\n    [[7, 4, 6, 5], [4, 7, 9, 2], [3, 1, 6, 5]],\n    [[4, 6, 8, 1], [11, 0, 3, 6], [0, 3, 6, 8]],\n    [[5, 3, 7, 5], [7, 7, 6, 2], [3, 8, 9, 2]]\n]\n\nobserved_data = [\n    [[8, 7, 11, 4], [6, 7, 6, 0], [3, 3, 7, 3]],\n    [[8, 8, 9, 5], [12, 1, 2, 8], [1, 4, 3, 12]],\n    [[5, 6, 6, 2], [3, 6, 4, 7], [5, 7, 8, 2]]\n]\n\n# Creating xarrays\nx = np.arange(3)\ny = np.arange(3)\ntime = pd.date_range('1/1/2011', periods=4, freq='D')\nforecast = xr.DataArray(\n    forecast_data,\n    coords=[('x', x), ('y', y), ('time', time)])\nobserved = xr.DataArray(\n    observed_data,\n    coords=[('x', x), ('y', y), ('time', time)])\n\n# Applying the contingency function\ncont = contingency(5, forecast, observed)\n\n# Generating skill metrics\npod = mt.pod(cont)\nfar = mt.far(cont)\ncsi = mt.csi(cont)\nfreq_bias = mt.freq_bias(cont)\naccuracy = mt.accuracy(cont)\npofd = mt.pofd(cont)\nhss = mt.hss(cont)\nets = mt.ets(cont)\nf1 = mt.f1Score(forecast, observed, average='weighted')\n\n# Displaying results as pandas series\nd = {\n    'Hits': cont[0],\n    'Misses': cont[1],\n    'False Alarm': cont[2],\n    'Correct Negative': cont[3],\n    'Probability of Detection': pod,\n    'False Alarm Ratio': far,\n    'Critical Success Index': csi,\n    'Frequency Bias': freq_bias,\n    'Accuracy': accuracy,\n    'Probability of False Detection': pofd,\n    'Heidke Skill Score': hss,\n    'Equitable Threat Score': ets,\n    'f1 Score': f1\n}\n\npd_series_d = pd.Series(d)\nprint(pd_series_d)"
      ]
    }
  ],
  "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": "3.6.15"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}