lesson_5_analysing_results.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lesson 5: Analysing results\n",
    "\n",
    "* Sanity Checks\n",
    "* Evaluating experiment: Single Evaluation Metric, Multiple Evaluation Metrics\n",
    "* Gotchas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy.stats import norm, binom, binom_test\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pylab import rcParams\n",
    "\n",
    "rcParams['figure.figsize'] = (10, 6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Sanity Checks - invariant metrics\n",
    "* Population sizing metrics  based on unit of diversion = Ensure that experiment and control populations are comparable\n",
    "* Any other invariant metrics = Check whether invariance metrics did not change during experiment\n",
    "\n",
    "### 1.1. Select Invariant metrics - based on unit of diversion!\n",
    "<img src=\"images/lesson5_quiz_selecting_invariants.png\">\n",
    "<img src=\"images/lesson5_quiz_selecting_invariants_2.png\">\n",
    "\n",
    "### 1.2 Check invariant metrics - populations sizing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bayesian approach"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from scipy.stats import binom\n",
    "test_split = (64454, 61818) # control, test\n",
    "lower = binom.ppf(n=sum(test_split), p=0.5, q=0.025) \n",
    "upper = binom.ppf(n=sum(test_split), p=0.5, q=0.975)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.48956221490116575"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_split[1] / sum(test_split)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(62788.0, 63484.0)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lower, upper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "61818"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_split[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.2113361931481954e-13"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "binom_test(test_split, p=0.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Frequentist approach"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.500587084148728"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_split = (15348, 15312) # control, test\n",
    "test_split[0] / sum(test_split)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.00285551160217\n"
     ]
    }
   ],
   "source": [
    "p = 0.5\n",
    "z = 1.96\n",
    "se = np.sqrt(p*(1-p)/sum(test_split))\n",
    "print(se)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.49440319725975251, 0.50559680274024754)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p - z * se, p + z * se"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### What to do if sanity checks failed\n",
    "\n",
    "* Check technical set-up with engineers\n",
    "* Retrospective analysis: try to reconstruct experiment diversion from previous data\n",
    "* Pre-/post-period. If I saw same change in pre\n",
    "\n",
    "Most common causes: \n",
    "* data capture\n",
    "* experiment set-up\n",
    "* infrastracture\n",
    "\n",
    "Possibly not a learning effect if we see big changes/differences right from the beginning."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Evaluating experiment\n",
    "\n",
    "## 2.1 Single evaluation metric\n",
    "\n",
    "Goal: make business decision based on whether the experiment favorably affected the summary metric.\n",
    "\n",
    "Analytically: decide if change in summary metric was statistically significant, also taking into cosideration magnitude and direction of change.\n",
    "\n",
    "\n",
    "**Is significant check** \n",
    "\n",
    "Combine results of previous steps\n",
    "\n",
    "* Characterising metrics, how it behaves (lesson 3)\n",
    "* Variability estimation => how long to run the experiment (lesson 4)\n",
    "\n",
    "\n",
    "**What if not significant?**\n",
    "\n",
    "* Look into days of week, slice by groups etc. to better understand the response and find some possible bugs.\n",
    "* Cross-check results of parametric vs. non-parametric (e.g. sign-test) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Quiz 9 \n",
    "https://classroom.udacity.com/courses/ud257/lessons/4085798776/concepts/40737188220923"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0116010196035\n"
     ]
    }
   ],
   "source": [
    "# data\n",
    "Xs_cont = np.array([196, 200, 200, 216, 212, 185, 225, 187, 205, 211, 192, 196, 223, 192])\n",
    "Ns_cont = np.array([2029, 1991, 1951, 1985, 1973, 2021, 2041, 1980, 1951, 1988, 1977, 2019, 2035, 2007])\n",
    "\n",
    "Xs_exp = np.array([179, 208, 205, 175, 191, 291, 278, 216, 225, 207, 205, 200, 297, 299])\n",
    "Ns_exp = np.array([1971, 2009, 2049, 2015, 2027, 1979, 1959, 2020, 2049, 2012, 2023, 1981, 1965, 1993])\n",
    "\n",
    "# meta\n",
    "d_min = 0.01\n",
    "alpha = 0.05\n",
    "z = round(norm.ppf(1-alpha/2), 2)\n",
    "\n",
    "size_empir = 5000\n",
    "se_empir = 0.0062 # empricial Standard Error with 5000 in each group\n",
    "\n",
    "\n",
    "dif = sum(Xs_exp) / sum(Ns_exp) - sum(Xs_cont)/sum(Ns_cont)\n",
    "print(dif)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$\\frac{\\text{se empir}}{\\text{scale empir}} = \\frac{\\text{SE}}{\\text{scal experim}}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.002619982707795742, 0.0062)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scal_experim = np.sqrt(1 / sum(Ns_cont) + 1/sum(Ns_exp))\n",
    "scal_empir = np.sqrt(1 / size_empir + 1/size_empir)\n",
    "se = se_empir * scal_experim / scal_empir\n",
    "se, se_empir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0064658534962369341, 0.016736185710796242)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "effect_size = (dif - z * se, dif + z * se)\n",
    "effect_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Ps_cont = Xs_cont / Ns_cont\n",
    "Ps_exp = Xs_exp / Ns_exp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.42395019531249989"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# p_value of sign test\n",
    "binom_test(x=sum(Ps_exp > Ps_cont), p=0.5, n=len(Ps_cont))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAF1CAYAAADSlV/tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJTNnoAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11053b9e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(Ps_cont).plot(label='control')\n",
    "pd.Series(Ps_exp).plot(label='experiment')\n",
    "_ = plt.legend()\n",
    "_ = plt.title('CTR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "control_share = Ns_cont / (Ns_cont + Ns_exp)\n",
    "total = Ns_cont + Ns_exp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2091.0"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "binom.ppf(n=total[0], p=control_share[0], q=0.975)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmMAAAFpCAYAAADQuy+GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIiH/j9eIb0Um97QxwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119cef1d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series(control_share).plot()\n",
    "_ = plt.axhline(0.5, ls='--', color='red')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Recommendation: stop and understand why CTR is so much higher on the weekends"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Single Metric Gotchas: Simpoon's paradox\n",
    "https://classroom.udacity.com/courses/ud257/lessons/4085798776/concepts/41091086150923"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"images/lesson5_simposons_paradox_ex.png\">\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# row: department A-B, columns men, women\n",
    "applied = np.array([[825, 108],\n",
    "                    [417, 375]])\n",
    "\n",
    "accepted = np.array([[512, 89],\n",
    "                     [137, 132]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.52254428,  0.45755694])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# acceptance rate per sex\n",
    "accepted.sum(axis=0) / applied.sum(axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.64415863,  0.33964646])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# acceptance rate per department\n",
    "accepted.sum(axis=1) / applied.sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.62060606,  0.82407407],\n",
       "       [ 0.32853717,  0.352     ]])"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# acceptance rate per department / sex\n",
    "accepted / applied "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.66425121,  0.22360248],\n",
       "       [ 0.33574879,  0.77639752]])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# share of men, women that applied to department a,b\n",
    "applied / applied.sum(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "77% of women applied to department with lower overall acceptance rate (35%)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2 Multiple evaluation metrics\n",
    "\n",
    "Testing more metrics, makes it more likely that one of them will show a statistically significant result by chance.\n",
    "https://en.wikipedia.org/wiki/Multiple_comparisons_problem\n",
    "\n",
    "Two main approaches:\n",
    "* control for familywise error rate\n",
    "* control for false discovery rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " probability of at least one metric being statistically significant out of 10 tested at 95%, 99% confidence interval"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.4012630607616213, 0.09561792499119559)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# assuming independence between metrics\n",
    "n = 10 # count null hypotheses\n",
    "alpha1 = 0.95\n",
    "alpha2 = 0.99\n",
    "1 - alpha1**n, 1 - alpha2**n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Control familywise error rate (FWER)\n",
    "https://en.wikipedia.org/wiki/Family-wise_error_rate\n",
    "\n",
    "$\\alpha_{familywise} = 1 - (1-\\alpha_{individual})^m$ \n",
    "\n",
    "without independence assumption:\n",
    "\n",
    "$\\alpha_{familywise} = 1 - m \\cdot \\alpha_{individual}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from multitesting_corr import p_multitest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Bonferroni correction \n",
    "https://en.wikipedia.org/wiki/Bonferroni_correction\n",
    "\n",
    "* +simple\n",
    "* +no assumptions\n",
    "* -too conservative if metrics are not independent or there are many experiments => reduces statistical power"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.05 0.0125\n",
      "1.95996398454 2.49770547441\n"
     ]
    }
   ],
   "source": [
    "# metrics are assumed to be normally distributed\n",
    "d_hat = np.array([0.03, -0.5, 0.01, 10])   # differences b/w test and control for each metric\n",
    "se = np.array([0.013, 0.21, 0.0045, 6.85]) # standard errors\n",
    "\n",
    "alpha_ind = 0.05\n",
    "alpha_bonf = alpha_ind / len(d_hat)\n",
    "z_ind = abs(norm.ppf(alpha_ind / 2)) # critical z-score\n",
    "z_bonf = abs(norm.ppf(alpha_bonf / 2))\n",
    "\n",
    "print(alpha_ind, alpha_bonf)\n",
    "print(z_ind, z_bonf)\n",
    "pvals = norm.pdf(d_hat / se)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True,  True,  True, False], dtype=bool)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abs(d_hat / se) > z_ind"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False, False, False], dtype=bool)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abs(d_hat / se) > z_bonf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0.02343778879550024: 0.093751155182000959,\n",
       " 0.027829451107324859: 0.11131780442929944,\n",
       " 0.033773651035270635: 0.13509460414108254,\n",
       " 0.13744582971982028: 0.54978331887928111}"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pvals_bonf = p_multitest(pvals, 'bonferroni')\n",
    "pvals_bonf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Holm-Bonferroni Method\n",
    "https://en.wikipedia.org/wiki/Holm%E2%80%93Bonferroni_method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pvals_bonf = p_multitest(pvals, 'bonferroni')\n",
    "pvals_holm = p_multitest(pvals, 'holm')\n",
    "pvals_bonf = pd.DataFrame.from_dict(pvals_bonf, orient='index').sort_index()\n",
    "pvals_holm = pd.DataFrame.from_dict(pvals_holm, orient='index').sort_index()\n",
    "methods_comparis = pd.concat([pvals_bonf, pvals_holm], axis=1)\n",
    "methods_comparis.columns = ['bonferroni', 'holm']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bonferroni</th>\n",
       "      <th>holm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.023438</th>\n",
       "      <td>0.093751</td>\n",
       "      <td>0.093751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.027829</th>\n",
       "      <td>0.111318</td>\n",
       "      <td>0.093751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.033774</th>\n",
       "      <td>0.135095</td>\n",
       "      <td>0.093751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.137446</th>\n",
       "      <td>0.549783</td>\n",
       "      <td>0.137446</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          bonferroni      holm\n",
       "0.023438    0.093751  0.093751\n",
       "0.027829    0.111318  0.093751\n",
       "0.033774    0.135095  0.093751\n",
       "0.137446    0.549783  0.137446"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "methods_comparis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmQAAAF3CAYAAAALu1cUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11a057dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "methods_comparis.plot(marker='o')\n",
    "_ = plt.xlabel('Original p-value')\n",
    "_ = plt.ylabel('Adjusted p-value')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tests\n",
    "http://www-stat.wharton.upenn.edu/~steele/Courses/956/Resource/MultipleComparision/Writght92.pdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0.0050000000000000001: 0.02,\n",
       " 0.01: 0.029999999999999999,\n",
       " 0.029999999999999999: 0.059999999999999998,\n",
       " 0.040000000000000001: 0.059999999999999998}"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Test 1\n",
    "test_pvals = np.array([0.01, 0.04, 0.03, 0.005])\n",
    "test_pvals_holm = p_multitest(test_pvals,'holm')\n",
    "test_pvals_holm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0.0029399999999999999: 0.017639999999999999,\n",
       " 0.0066800000000000002: 0.040080000000000005,\n",
       " 0.069400000000000003: 0.41639999999999999,\n",
       " 0.13902999999999999: 0.83417999999999992,\n",
       " 0.70221999999999996: 1}"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Test 2\n",
    "test_pvals = np.array([0.00294, 0.00668, 0.06940, 0.13903, 0.13903, 0.70222])\n",
    "test_pvals_bonf = p_multitest(test_pvals,'bonferroni')\n",
    "test_pvals_bonf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0.0029399999999999999: 0.017639999999999999,\n",
       " 0.0066800000000000002: 0.033399999999999999,\n",
       " 0.069400000000000003: 0.27760000000000001,\n",
       " 0.13902999999999999: 0.41708999999999996,\n",
       " 0.70221999999999996: 0.70221999999999996}"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Test 2\n",
    "test_pvals_holm = p_multitest(test_pvals,'holm')\n",
    "test_pvals_holm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0.0029399999999999999: 0.017639999999999999,\n",
       " 0.0066800000000000002: 0.020040000000000002,\n",
       " 0.069400000000000003: 0.13880000000000001,\n",
       " 0.13902999999999999: 0.16683599999999998,\n",
       " 0.70221999999999996: 0.70221999999999996}"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Test 3\n",
    "test_pvals_fdr = p_multitest(test_pvals, 'fdr')\n",
    "test_pvals_fdr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pvals_bonf = pd.DataFrame.from_dict(test_pvals_bonf, orient='index').sort_index()\n",
    "pvals_holm = pd.DataFrame.from_dict(test_pvals_holm, orient='index').sort_index()\n",
    "pvals_fdr = pd.DataFrame.from_dict(test_pvals_fdr, orient='index').sort_index()\n",
    "methods_comparis = pd.concat([pvals_bonf, pvals_holm, pvals_fdr], axis=1)\n",
    "methods_comparis.columns = ['bonferroni', 'holm', 'fdr']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmcAAAF3CAYAAADgjOwXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVHUhPQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11a237400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha = 0.05\n",
    "methods_comparis.plot(marker='o')\n",
    "_ = plt.axhline(alpha, color='red', ls='--')\n",
    "_ = plt.xlabel('Original p-value')\n",
    "_ = plt.ylabel('Adjusted p-value')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Resampling procedures"
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}