2_results.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results\n",
    "\n",
    "This notebook shows how to work with the results from the joint Crab paper.\n",
    "\n",
    "We will give a quick overview of the files in the `results` folder, and then show how to load the spectral fit results (best-fit parameters and covariance matrix) and make a plot of the parameter constraints.\n",
    "\n",
    "As explained in the [README.md](README.md), the main analysis and the figures for this paper were produced by Python scripts (`make.py` and `joint_crab` package). For further examples, please look there."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from multinorm import MultiNorm\n",
    "from astropy.table import Table"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Overview\n",
    "\n",
    "All results are stored in standard machine-readable formats that are easy to read from Python. We used the YAML format for structured data (using the Python `yaml` package), ECSV and FITS for table (using the Python `astropy` package)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "README.md\n",
      "\u001b[1m\u001b[34merrorbands\u001b[m\u001b[m\n",
      "\u001b[1m\u001b[34mfigures\u001b[m\u001b[m\n",
      "\u001b[1m\u001b[34mfit\u001b[m\u001b[m\n",
      "\u001b[1m\u001b[34mmaps\u001b[m\u001b[m\n",
      "provenance.yaml\n",
      "\u001b[1m\u001b[34mspectra\u001b[m\u001b[m\n",
      "\u001b[1m\u001b[34msummary\u001b[m\u001b[m\n"
     ]
    }
   ],
   "source": [
    "!ls -1 results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "contours_fact.yaml\n",
      "contours_fermi.yaml\n",
      "contours_hess.yaml\n",
      "contours_joint.yaml\n",
      "contours_joint_systematics.yaml\n",
      "contours_magic.yaml\n",
      "contours_veritas.yaml\n",
      "fit_fact.yaml\n",
      "fit_fermi.yaml\n",
      "fit_hess.yaml\n",
      "fit_joint.yaml\n",
      "fit_joint_systematics.yaml\n",
      "fit_magic.yaml\n",
      "fit_veritas.yaml\n"
     ]
    }
   ],
   "source": [
    "!ls -1 results/fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "samples_fact.fits.gz\n",
      "samples_fermi.fits.gz\n",
      "samples_hess.fits.gz\n",
      "samples_joint.fits.gz\n",
      "samples_magic.fits.gz\n",
      "samples_veritas.fits.gz\n",
      "sed_fact.ecsv\n",
      "sed_fermi.ecsv\n",
      "sed_hess.ecsv\n",
      "sed_joint.ecsv\n",
      "sed_magic.ecsv\n",
      "sed_veritas.ecsv\n"
     ]
    }
   ],
   "source": [
    "!ls -1 results/errorbands"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data.md\n",
      "data.tex\n",
      "results.md\n",
      "results.tex\n"
     ]
    }
   ],
   "source": [
    "!ls -1 results/summary/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read YAML fit results\n",
    "\n",
    "The fit results are given in the files `results/fit/fit_{name}.yaml`, where `name = {\"fermi\", \"magic\", \"veritas\", \"fact\", \"hess\", \"joint\"}`. As an example, you can look at [results/fit/fit_fermi.yaml](results/fit/fit_fermi.yaml) to see the structure. For each fit parameter, there's the name, value and error, and there's also the covariance matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'parameters': [{'name': 'amplitude',\n",
       "   'value': 3.8491280706505666e-11,\n",
       "   'factor': 3.8491280706505666,\n",
       "   'scale': '1e-11',\n",
       "   'unit': 'cm-2 s-1 TeV-1',\n",
       "   'min': nan,\n",
       "   'max': nan,\n",
       "   'frozen': False,\n",
       "   'error': 1.1435170001020147e-12},\n",
       "  {'name': 'reference',\n",
       "   'value': 1.0,\n",
       "   'factor': 1.0,\n",
       "   'scale': 1.0,\n",
       "   'unit': 'TeV',\n",
       "   'min': nan,\n",
       "   'max': nan,\n",
       "   'frozen': True,\n",
       "   'error': 0.0},\n",
       "  {'name': 'alpha',\n",
       "   'value': 2.5053735278333207,\n",
       "   'factor': 2.5053735278333207,\n",
       "   'scale': 1.0,\n",
       "   'unit': '',\n",
       "   'min': nan,\n",
       "   'max': nan,\n",
       "   'frozen': False,\n",
       "   'error': 0.02549508100457278},\n",
       "  {'name': 'beta',\n",
       "   'value': 0.2362106123242955,\n",
       "   'factor': 2.3621061232429548,\n",
       "   'scale': 0.1,\n",
       "   'unit': '',\n",
       "   'min': nan,\n",
       "   'max': nan,\n",
       "   'frozen': False,\n",
       "   'error': 0.02450886256203386}],\n",
       " 'covariance': [[1.3076311295223112e-24,\n",
       "   0.0,\n",
       "   1.2371922299026042e-16,\n",
       "   1.1603842009321198e-14],\n",
       "  [0.0, 0.0, 0.0, 0.0],\n",
       "  [1.2371922299026042e-16, 0.0, 0.0006499991554297277, 0.000492142818491591],\n",
       "  [1.1603842009321198e-14, 0.0, 0.000492142818491591, 0.0006006843440846649]],\n",
       " 'statname': 'wstat',\n",
       " 'statval': 41.12925420683606,\n",
       " 'fit_range': {'min': 891250938.1337459,\n",
       "  'max': 28183829312.64455,\n",
       "  'unit': 'keV'}}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "info = yaml.load(open(\"results/fit/fit_joint.yaml\"))\n",
    "info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Usually you'll have to write a few lines of code to get the data in the form you want.\n",
    "\n",
    "For example, here's how to put the parameter information into a `pandas.DataFrame`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>name</th>\n",
       "      <th>value</th>\n",
       "      <th>error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>amplitude</td>\n",
       "      <td>3.849128e-11</td>\n",
       "      <td>1.143517e-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>reference</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>alpha</td>\n",
       "      <td>2.505374e+00</td>\n",
       "      <td>2.549508e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>beta</td>\n",
       "      <td>2.362106e-01</td>\n",
       "      <td>2.450886e-02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        name         value         error\n",
       "0  amplitude  3.849128e-11  1.143517e-12\n",
       "1  reference  1.000000e+00  0.000000e+00\n",
       "2      alpha  2.505374e+00  2.549508e-02\n",
       "3       beta  2.362106e-01  2.450886e-02"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(\n",
    "    [\n",
    "        {\"name\": _[\"name\"], \"value\": _[\"value\"], \"error\": _[\"error\"]}\n",
    "        for _ in info[\"parameters\"]\n",
    "    ],\n",
    "    columns=[\"name\", \"value\", \"error\"],\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `yaml.load` call returns Python objects (str, float, list, dict). Often for computations you want a Numpy array. It's easy to make:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.30763113e-24, 0.00000000e+00, 1.23719223e-16, 1.16038420e-14],\n",
       "       [0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00],\n",
       "       [1.23719223e-16, 0.00000000e+00, 6.49999155e-04, 4.92142818e-04],\n",
       "       [1.16038420e-14, 0.00000000e+00, 4.92142818e-04, 6.00684344e-04]])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cov = np.array(info[\"covariance\"])\n",
    "cov"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check parameter correlations\n",
    "\n",
    "If you look at Table 2 or Figure 4 in the paper, you will find that the parameter errors in the joint fit are much smaller than the parameter errors from each individual instrument dataset.\n",
    "\n",
    "Why is that?\n",
    "\n",
    "The reason is that there are strong correlations between the three fit parameters, especially for the Fermi-LAT dataset. The likelihood for each dataset is approximately a three-dimensional multivariate normal, and the joint likelihood is approximately the product of those likelihood, and the parameter correlations lead to a well-constrained joint likelihood.\n",
    "\n",
    "Let's load all fit results and compute the product multivariate normal, and then create a Figure similar to Fig. 4 in the paper. We will use the `MultiNorm` class from the [multinorm](https://multinorm.readthedocs.io/) module to help a bit with the covariance matrix related computations.\n",
    "\n",
    "### Load all results\n",
    "\n",
    "First let's load all fit results, and store each one in a `MultiNorm` object:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_fit_result(name):\n",
    "    data = yaml.load(open(f\"results/fit/fit_{name}.yaml\"))\n",
    "    mean = [_[\"value\"] for _ in data[\"parameters\"]]\n",
    "    cov = data[\"covariance\"]\n",
    "    # Use parameter names that are consistent with the paper\n",
    "    names = [\"phi\", \"reference\", \"gamma\", \"beta\"]\n",
    "    mn = MultiNorm(mean, cov, names)\n",
    "    mn = mn.drop([\"reference\"])\n",
    "    return mn\n",
    "\n",
    "\n",
    "results = {\n",
    "    name: read_fit_result(name)\n",
    "    for name in (\"fermi\", \"magic\", \"veritas\", \"fact\", \"hess\", \"joint\")\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check results\n",
    "\n",
    "We can now do any covariance-matrix related computations, e.g. to compute parameter errors or correlations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>mean</th>\n",
       "      <th>err</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>phi</th>\n",
       "      <td>3.849128e-11</td>\n",
       "      <td>1.143517e-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gamma</th>\n",
       "      <td>2.505374e+00</td>\n",
       "      <td>2.549508e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta</th>\n",
       "      <td>2.362106e-01</td>\n",
       "      <td>2.450886e-02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               mean           err\n",
       "name                             \n",
       "phi    3.849128e-11  1.143517e-12\n",
       "gamma  2.505374e+00  2.549508e-02\n",
       "beta   2.362106e-01  2.450886e-02"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# These are the numbers in the last row in Table 2 in the paper\n",
    "results[\"joint\"].parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>name</th>\n",
       "      <th>phi</th>\n",
       "      <th>gamma</th>\n",
       "      <th>beta</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>phi</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.910732</td>\n",
       "      <td>-0.797599</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gamma</th>\n",
       "      <td>-0.910732</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.972449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta</th>\n",
       "      <td>-0.797599</td>\n",
       "      <td>0.972449</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name        phi     gamma      beta\n",
       "name                               \n",
       "phi    1.000000 -0.910732 -0.797599\n",
       "gamma -0.910732  1.000000  0.972449\n",
       "beta  -0.797599  0.972449  1.000000"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The parameter correlations for Fermi-LAT are very high\n",
    "results[\"fermi\"].correlation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.9107317393899312"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[\"fermi\"].correlation[\"phi\"][\"gamma\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Product likelihood\n",
    "\n",
    "Let's compute the product multivariate normal, which is an approximation for the joint fit result."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_joint_product_result(results):\n",
    "    labels = [\"fermi\", \"magic\", \"veritas\", \"fact\", \"hess\"]\n",
    "    distributions = [results[_] for _ in labels]\n",
    "    mn = MultiNorm.from_product(distributions)\n",
    "    return mn\n",
    "\n",
    "\n",
    "results[\"product\"] = compute_joint_product_result(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>mean</th>\n",
       "      <th>err</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>phi</th>\n",
       "      <td>3.828777e-11</td>\n",
       "      <td>1.188333e-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gamma</th>\n",
       "      <td>2.489757e+00</td>\n",
       "      <td>2.765209e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta</th>\n",
       "      <td>2.203237e-01</td>\n",
       "      <td>2.569859e-02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               mean           err\n",
       "name                             \n",
       "phi    3.828777e-11  1.188333e-12\n",
       "gamma  2.489757e+00  2.765209e-02\n",
       "beta   2.203237e-01  2.569859e-02"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[\"product\"].parameters"
   ]
  },
  {
   "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>mean</th>\n",
       "      <th>err</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>phi</th>\n",
       "      <td>3.849128e-11</td>\n",
       "      <td>1.143517e-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gamma</th>\n",
       "      <td>2.505374e+00</td>\n",
       "      <td>2.549508e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta</th>\n",
       "      <td>2.362106e-01</td>\n",
       "      <td>2.450886e-02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               mean           err\n",
       "name                             \n",
       "phi    3.849128e-11  1.143517e-12\n",
       "gamma  2.505374e+00  2.549508e-02\n",
       "beta   2.362106e-01  2.450886e-02"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[\"joint\"].parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot\n",
    "\n",
    "Let's plot the 2-dimensional parameter contraints (\"1 sigma error ellipses\") for each pair of parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_ellipses_panel(results, names, ax, legend):\n",
    "    colors = [\n",
    "        \"#21ABCD\",\n",
    "        \"#FF9933\",\n",
    "        \"#893F45\",\n",
    "        \"#3EB489\",\n",
    "        \"#002E63\",\n",
    "        \"crimson\",\n",
    "        \"black\",\n",
    "    ]\n",
    "    for name, color in zip(results, colors):\n",
    "        m = results[name].marginal(names)\n",
    "        ellipse = m.to_matplotlib_ellipse(ec=color, fc=\"none\", label=name)\n",
    "        ax.add_patch(ellipse)\n",
    "        ax.plot(m.mean[0], m.mean[1], \".\", color=color)\n",
    "    ax.set_xlabel(names[0])\n",
    "    ax.set_ylabel(names[1])\n",
    "    if legend:\n",
    "        ax.legend(loc=1)\n",
    "\n",
    "\n",
    "def plot_ellipses(results):\n",
    "    fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(10, 3))\n",
    "    plot_ellipses_panel(results, [\"phi\", \"gamma\"], axes[0], legend=False)\n",
    "    plot_ellipses_panel(results, [\"phi\", \"beta\"], axes[1], legend=False)\n",
    "    plot_ellipses_panel(results, [\"gamma\", \"beta\"], axes[2], legend=True)\n",
    "    fig.tight_layout()\n",
    "\n",
    "\n",
    "plot_ellipses(results)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As mentioned at the start, we can see that the joint fit result is approximately given by the product of the multivariate normal approximation for the fit result of each individual dataset in this analysis."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SED\n",
    "\n",
    "In `results/errorbands`, the SEDs corresponding to the best-fit model are pre-computed, including 1 sigma error bands. Here we show how to upen up such a table and plot the SED."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "sed = Table.read(\"results/errorbands/sed_veritas.ecsv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<Table length=81>\n",
      "     name      dtype        unit      format\n",
      "------------- ------- --------------- ------\n",
      "       energy float64             TeV     3g\n",
      "    dnde_mean float64 1 / (cm2 s TeV)     3g\n",
      "  dnde_median float64 1 / (cm2 s TeV)     3g\n",
      "      dnde_lo float64 1 / (cm2 s TeV)     3g\n",
      "      dnde_hi float64 1 / (cm2 s TeV)     3g\n",
      "     dnde_fit float64 1 / (cm2 s TeV)     3g\n",
      "   e2dnde_fit float64   erg / (cm2 s)     3g\n",
      "  e2dnde_mean float64   erg / (cm2 s)     3g\n",
      "e2dnde_median float64   erg / (cm2 s)     3g\n",
      "    e2dnde_lo float64   erg / (cm2 s)     3g\n",
      "    e2dnde_hi float64   erg / (cm2 s)     3g\n"
     ]
    }
   ],
   "source": [
    "sed.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'E^2 (dN/dE) (erg cm-2 s-1)')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x120524e10>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.plot(sed[\"energy\"], sed[\"e2dnde_fit\"], lw=2, c=\"r\")\n",
    "ax.fill_between(sed[\"energy\"], sed[\"e2dnde_lo\"], sed[\"e2dnde_hi\"], alpha=0.1)\n",
    "ax.loglog()\n",
    "ax.set_xlabel(\"Energy (TeV)\")\n",
    "ax.set_ylabel(\"E^2 (dN/dE) (erg cm-2 s-1)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some key points to take away:\n",
    "\n",
    "- YAML is great to store structured data, and easy to read / write with the Python `yaml` module.\n",
    "- ECSV is nice to store tabular data if you use `astropy.table.Table`.\n",
    "- FITS is also a good format for tabular data. It supports columns which contain multidimensional arrays, which ECSV does not.\n",
    "- Separate computation from plotting, for easier reproducibility, debugging and re-use."
   ]
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}