Drift_Fourier_Analysis_Bootstrap_CCL4_NYA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Trend analysis with bootstrap resampling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Python / Numpy implementation of the trend analysis presented in [Gardiner et al., 2008][1]\n",
    "\n",
    "The following model is used to fit the annual trend (drift) + intra-annual variability (i.e., fourier series tuncated at $n$ degrees):\n",
    "\n",
    "$$ F(t, p, \\alpha, \\mathrm{{\\mathbf \\beta}}) = p + \\alpha t + \\left[\\beta_1 \\cos(\\frac{\\pi t}{L}) + \\beta_2 \\sin(\\frac{\\pi t}{L}) + \\beta_3 \\cos(\\frac{2 \\pi t}{L}) + \\beta_4 \\sin(\\frac{2 \\pi t}{L}) + \\cdots +  \\beta_{2n-1} \\cos(\\frac{n \\pi t}{L}) + \\beta_{2n} \\sin(\\frac{n \\pi t}{L}) \\right]$$\n",
    "\n",
    "where $p$ is the intercept, $\\alpha$ is the trend (slope), $\\mathrm{{\\mathbf \\beta}}$ are the truncated fourier series parameters, and $L$ is the half-period (= 0.5 if $t$ is in years).\n",
    "\n",
    "Model parameters are estimated using the weighted linear least squares method. Bootstrap resampling can be used to evaluate the confidence intervals on the parameter estimates (more robust in case of non-normally distributed random effects in the data).\n",
    "\n",
    "Author: B. Bovy | GIRPAS, ULg\n",
    "\n",
    "[1]: http://dx.doi.org/10.5194/acp-8-6719-2008"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "%matplotlib inline\n",
    "sns.set_context('notebook')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Implementation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def fourier_basis(x, degree, half_period):\n",
    "    \"\"\"Returns a 2-d array of fourier basis.\"\"\"\n",
    "    A = np.ones((x.size, 2 * degree + 1))\n",
    "    \n",
    "    for d in range(1, degree + 1):\n",
    "        A[:, 2*d-1] = np.cos(d * np.pi * x / half_period)\n",
    "        A[:, 2*d] = np.sin(d * np.pi * x / half_period)\n",
    "    \n",
    "    return A\n",
    "\n",
    "\n",
    "def fit_driftfourier(x, data, weights, degree, half_period=0.5):\n",
    "    \"\"\"\n",
    "    Fit y = f(x - x.min()) to data where f is given by\n",
    "    fourier series + drift.\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    x : 1-d array\n",
    "        x-coordinates\n",
    "    data : 1-d array\n",
    "        data values\n",
    "    weights : 1-d array\n",
    "        weights (>=0)\n",
    "    degree : int\n",
    "        degree of fourier series\n",
    "    half_period : float\n",
    "        half period\n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "    intercept : float\n",
    "        intercept at x.min()\n",
    "    slope : float\n",
    "        slope (drift) for the normalized data\n",
    "        (x - x.min())\n",
    "    pfourier : 1-d array\n",
    "        Fourier series parameters for the\n",
    "        normalized data\n",
    "    f_drift : callable\n",
    "        Can be used to calculate the drift\n",
    "        given any (non-normalized) x\n",
    "    f_fourier : callable\n",
    "        Can be used to calculate fourier series\n",
    "    f_driftfourier : callable\n",
    "        Can be used to calculate drift + fourier\n",
    "    residual_std : float\n",
    "        estimated standard deviation of residuals\n",
    "    A : 2-d array\n",
    "        matrix of \"coefficients\"\n",
    "    \n",
    "    \"\"\"\n",
    "    xmin = x.min()\n",
    "    xnorm = x - xmin\n",
    "    \n",
    "    # coefficient matrix\n",
    "    A = np.ones((x.size, 2 * degree + 2))\n",
    "    A[:, 1] = xnorm\n",
    "    A[:, 2:] = fourier_basis(xnorm, degree, half_period)[:, 1:]\n",
    "    \n",
    "    # linear weighted least squares\n",
    "    results = np.linalg.lstsq(A * weights[:, np.newaxis],\n",
    "                              data * weights)\n",
    "    \n",
    "    params = results[0]\n",
    "    intercept = params[0]\n",
    "    slope = params[1]\n",
    "    pfourier = params[2:]\n",
    "    \n",
    "    f_drift = lambda t: slope * (t - xmin) + intercept\n",
    "    f_fourier = lambda t: np.sum(fourier_basis(t - xmin, degree,\n",
    "                                               half_period)[:, 1:]\n",
    "                                 * pfourier[np.newaxis, :],\n",
    "                                 axis=1) + intercept\n",
    "    f_driftfourier = lambda t: f_drift(t) + f_fourier(t) - intercept\n",
    "    \n",
    "    residual_std = np.sqrt(results[1][0] / (x.size - 2 * degree + 2)) \n",
    "    \n",
    "    return (intercept, slope, pfourier,\n",
    "            f_drift, f_fourier, f_driftfourier,\n",
    "            residual_std, A)\n",
    "\n",
    "\n",
    "def cf_driftfourier(x, data, weights, degree,\n",
    "                    half_period=0.5, nboot=5000,\n",
    "                    percentiles=(2.5, 50., 97.5)):\n",
    "    \"\"\"\n",
    "    Calculate confidence intervals for the fitted\n",
    "    parameters from fourier series + drift modelling,\n",
    "    using bootstrap resampling.\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    nboot : int\n",
    "        number of bootstrap replicates\n",
    "    percentiles : sequence of floats\n",
    "        percentiles of parameter estimate\n",
    "        distributions to return \n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "    perc : dict\n",
    "        percentiles for each parameter\n",
    "        distribution.\n",
    "    boot_intercept : 1-d array\n",
    "        intercept estimates from bootstrapped\n",
    "        datasets.\n",
    "    boot_slope : 1-d array\n",
    "        slope estimates.\n",
    "    boot_pfourier : 2-d array\n",
    "        fourier parameters estimates.\n",
    "    \n",
    "    See Also\n",
    "    --------\n",
    "    :func:`fit_driftfourier`\n",
    "    \n",
    "    \"\"\"\n",
    "    # first fit without bootstrapping\n",
    "    results = fit_driftfourier(x, data, weights,\n",
    "                               degree, half_period)\n",
    "    f_driftfourier = results[5]\n",
    "    A = results[7]\n",
    "    model = f_driftfourier(x)\n",
    "    residuals = data - model\n",
    "    \n",
    "    # generate bootstrap resamples of residuals\n",
    "    # and new datasets from these resamples\n",
    "    resample_residuals = np.random.choice(\n",
    "        residuals, size=(x.size, nboot), replace=True\n",
    "    )\n",
    "    boot_dataset = model[:, np.newaxis] + resample_residuals\n",
    "    \n",
    "    # fit all bootstrap datasets at once\n",
    "    results_boot = np.linalg.lstsq(A * weights[:, np.newaxis],\n",
    "                                   boot_dataset * weights[:, np.newaxis])\n",
    "    \n",
    "    params_boot = results_boot[0]\n",
    "    \n",
    "    # compute percentiles\n",
    "    perc_boot = np.column_stack(\n",
    "        np.percentile(params_boot, percentiles, axis=1)\n",
    "    )\n",
    "    \n",
    "    perc = {'intercept' : perc_boot[0],\n",
    "            'slope' : perc_boot[1],\n",
    "            'pfourier' : perc_boot[2:]}\n",
    "    \n",
    "    intercept = params_boot[0]\n",
    "    slope = params_boot[1]\n",
    "    pfourier = params_boot[2:]\n",
    "    \n",
    "    return perc, intercept, slope, pfourier\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_slope_percent(slope, intercept):\n",
    "    \"\"\"\n",
    "    Express `slope` as percentage, taking\n",
    "    the `intercept` value as reference (100%).\n",
    "    \"\"\"\n",
    "    return slope / intercept * 100."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## CCL4 NY4 data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import the data from CSV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ccl4_nya = pd.read_csv('data/CCL4_NYA.csv', header=None,\n",
    "                       names=['time', 'col_density'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Fit trend only (no fourier parameters added to the model), without bootstrapping**. Also express trend as percentage."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "x = ccl4_nya['time'].values\n",
    "y = ccl4_nya['col_density'].values\n",
    "w = np.ones_like(x)\n",
    "fourier_degree = 0   # fit trend only\n",
    "\n",
    "res = fit_driftfourier(x, y, w, fourier_degree)\n",
    "intercept, slope, pfourier, residuals_std = res[0:3] + (res[6],)\n",
    "f_drift, f_fourier, f_driftfourier = res[3:6]\n",
    "slope_percent = get_slope_percent(slope, intercept)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "fitted trend: -2.018E+13 molec./cm2/yr | -1.083 %/yr\n",
      "fitted intercept at xmin: 1.863E+15 molec./cm2\n",
      "fitted fourier parameters: []\n",
      "std of residuals: 3.186E+14 molec./cm2\n"
     ]
    }
   ],
   "source": [
    "print(\"fitted trend: {:.3E} molec./cm2/yr | {:.3f} %/yr\".format(slope, slope_percent))\n",
    "print(\"fitted intercept at xmin: {:.3E} molec./cm2\".format(intercept))\n",
    "print(\"fitted fourier parameters: {}\".format(pfourier))\n",
    "print(\"std of residuals: {:.3E} molec./cm2\".format(residuals_std))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot data + fitted line"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f8f880253c8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/python/miniconda3/envs/main_py34/lib/python3.4/site-packages/matplotlib/collections.py:590: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  if self._edgecolors == str('face'):\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf4AAAFsCAYAAAAtwdttAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8f8a0ff8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ccl4_nya.plot(x='time', y='col_density', kind='scatter', label='data')\n",
    "plt.plot(x, f_drift(x), label='fitted annual trend')\n",
    "plt.plot(x, f_driftfourier(x),\n",
    "         label='fitted annual trend + intra-annual variability')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Fit trend only, with bootstrap resampling**. Compute 2.5% and 97.5% percentiles (i.e., the 95% interval) for each fitted parameter. Compute also the percentiles for trend in %."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ci95_percentiles = (2.5, 97.5)\n",
    "\n",
    "cf, boot_intercept, boot_slope, boot_pfourier = cf_driftfourier(\n",
    "    x, y, w, fourier_degree, nboot=20000,\n",
    "    percentiles=ci95_percentiles\n",
    ")\n",
    "\n",
    "boot_slope_percent = get_slope_percent(boot_slope, boot_intercept)\n",
    "cf['slope_percent'] = np.percentile(boot_slope_percent, ci95_percentiles)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "95% CI for trend: (-2.510E+13, -1.526E+13) molec./cm2/yr\n",
      "95% CI for trend: (-1.320, -0.836) %/yr\n",
      "95% CI for intercept: (1.817E+15, 1.909E+15) molec./cm2\n",
      "percentiles for fourier parameters: []\n"
     ]
    }
   ],
   "source": [
    "print(\"95% CI for trend: ({:.3E}, {:.3E}) molec./cm2/yr\".format(*cf['slope']))\n",
    "print(\"95% CI for trend: ({:.3f}, {:.3f}) %/yr\".format(*cf['slope_percent']))\n",
    "print(\"95% CI for intercept: ({:.3E}, {:.3E}) molec./cm2\".format(*cf['intercept']))\n",
    "print(\"percentiles for fourier parameters: {}\".format(cf['pfourier']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Histograms of the distributions of the annual trend and intercept estimates obtained from bootstrap resampling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvkAAADVCAYAAADNTe8cAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZJklPF/cPIc/vKDo/kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8f8800ecc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 3, figsize=(13, 3))\n",
    "\n",
    "def plot_bootdist(v, ax, label):\n",
    "    sns.distributions.distplot(v, ax=ax)\n",
    "    plt.setp(ax, xlabel=label, yticks=[])\n",
    "\n",
    "plot_bootdist(boot_slope, axes[0], 'trend (molec./cm2/yr)')\n",
    "plot_bootdist(boot_slope_percent, axes[1], 'trend (%/yr)')\n",
    "plot_bootdist(boot_intercept, axes[2], 'intercept (molec./cm2)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Main (py34)",
   "language": "python",
   "name": "main_py34"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.4.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}