{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Risk Analysis of the Space Shuttle: Pre-Challenger Prediction of Failure" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this document we reperform some of the analysis provided in \n", "*Risk Analysis of the Space Shuttle: Pre-Challenger Prediction of Failure* by *Siddhartha R. Dalal, Edward B. Fowlkes, Bruce Hoadley* published in *Journal of the American Statistical Association*, Vol. 84, No. 408 (Dec., 1989), pp. 945-957 and available at http://www.jstor.org/stable/2290069. \n", "\n", "On the fourth page of this article, they indicate that the maximum likelihood estimates of the logistic regression using only temperature are: $\\hat{\\alpha}=5.085$ and $\\hat{\\beta}=-0.1156$ and their asymptotic standard errors are $s_{\\hat{\\alpha}}=3.052$ and $s_{\\hat{\\beta}}=0.047$. The Goodness of fit indicated for this model was $G^2=18.086$ with 21 degrees of freedom. Our goal is to reproduce the computation behind these values and the Figure 4 of this article, possibly in a nicer looking way." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Technical information on the computer on which the analysis is run" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will be using the python3 language using the pandas, statsmodels, numpy, matplotlib and seaborn libraries." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "#!pip install statsmodels \n", "#!pip install seaborn" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "3.7.6 (default, Jan 8 2020, 19:59:22) \n", "[GCC 7.3.0]\n", "uname_result(system='Linux', node='H2-SCHMIDER', release='5.3.0-46-generic', version='#38~18.04.1-Ubuntu SMP Tue Mar 31 04:17:56 UTC 2020', machine='x86_64', processor='x86_64')\n", "IPython 7.13.0\n", "IPython.core.release 7.13.0\n", "html 6.0.3\n", "_csv 1.0\n", "_ctypes 1.1.0\n", "_curses b'2.2'\n", "decimal 1.70\n", "argparse 1.1\n", "backcall 0.1.0\n", "csv 1.0\n", "ctypes 1.1.0\n", "cycler 0.10.0\n", "dateutil 2.8.1\n", "decimal 1.70\n", "decorator 4.4.2\n", "distutils 3.7.6\n", "ipykernel 5.1.4\n", "ipykernel._version 5.1.4\n", "ipython_genutils 0.2.0\n", "ipython_genutils._version 0.2.0\n", "jedi 0.16.0\n", "json 2.0.9\n", "jupyter_client 6.1.2\n", "jupyter_client._version 6.1.2\n", "jupyter_core 4.6.3\n", "jupyter_core.version 4.6.3\n", "kiwisolver 1.2.0\n", "logging 0.5.1.2\n", "matplotlib 3.2.1\n", "matplotlib.backends.backend_agg 3.2.1\n", "numpy 1.18.2\n", "numpy.core 1.18.2\n", "numpy.core._multiarray_umath 3.1\n", "numpy.lib 1.18.2\n", "numpy.linalg._umath_linalg b'0.1.5'\n", "pandas 1.0.3\n", "parso 0.6.2\n", "patsy 0.5.1\n", "patsy.version 0.5.1\n", "pexpect 4.8.0\n", "pickleshare 0.7.5\n", "platform 1.0.8\n", "prompt_toolkit 3.0.4\n", "ptyprocess 0.6.0\n", "pygments 2.6.1\n", "pyparsing 2.4.6\n", "pytz 2019.3\n", "re 2.2.1\n", "scipy 1.4.1\n", "scipy._lib._uarray 0.5.1+5.ga864a57.scipy\n", "scipy._lib.decorator 4.0.5\n", "scipy._lib.six 1.2.0\n", "scipy.integrate._dop b'$Revision: $'\n", "scipy.integrate._ode $Id$\n", "scipy.integrate._odepack 1.9 \n", "scipy.integrate._quadpack 1.13 \n", "scipy.integrate.lsoda b'$Revision: $'\n", "scipy.integrate.vode b'$Revision: $'\n", "scipy.interpolate._fitpack 1.7 \n", "scipy.interpolate.dfitpack b'$Revision: $'\n", "scipy.linalg 0.4.9\n", "scipy.linalg._fblas b'$Revision: $'\n", "scipy.linalg._flapack b'$Revision: $'\n", "scipy.linalg._flinalg b'$Revision: $'\n", "scipy.ndimage 2.0\n", "scipy.optimize._cobyla b'$Revision: $'\n", "scipy.optimize._lbfgsb b'$Revision: $'\n", "scipy.optimize._minpack 1.10 \n", "scipy.optimize._nnls b'$Revision: $'\n", "scipy.optimize._slsqp b'$Revision: $'\n", "scipy.optimize.minpack2 b'$Revision: $'\n", "scipy.signal.spline 0.2\n", "scipy.sparse.linalg.eigen.arpack._arpack b'$Revision: $'\n", "scipy.sparse.linalg.isolve._iterative b'$Revision: $'\n", "scipy.special.specfun b'$Revision: $'\n", "scipy.stats.mvn b'$Revision: $'\n", "scipy.stats.statlib b'$Revision: $'\n", "seaborn 0.10.0\n", "seaborn.external.husl 2.1.0\n", "six 1.14.0\n", "statsmodels 0.11.1\n", "statsmodels.__init__ 0.11.1\n", "statsmodels.api 0.11.1\n", "statsmodels.tools.web 0.11.1\n", "traitlets 4.3.3\n", "traitlets._version 4.3.3\n", "urllib.request 3.7\n", "zlib 1.0\n", "zmq 18.1.1\n", "zmq.sugar 18.1.1\n", "zmq.sugar.version 18.1.1\n" ] } ], "source": [ "def print_imported_modules():\n", " import sys\n", " for name, val in sorted(sys.modules.items()):\n", " if(hasattr(val, '__version__')): \n", " print(val.__name__, val.__version__)\n", "# else:\n", "# print(val.__name__, \"(unknown version)\")\n", "def print_sys_info():\n", " import sys\n", " import platform\n", " print(sys.version)\n", " print(platform.uname())\n", "\n", "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "import statsmodels.api as sm\n", "import seaborn as sns\n", "\n", "print_sys_info()\n", "print_imported_modules()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Loading and inspecting data\n", "Let's start by reading data." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "'/home/aschmide/Documents/formation_MOOC_RR/mooc-rr/module4'" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pwd" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "file downloaded on Sun Apr 12 17:22:23 2020\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
DateCountTemperaturePressureMalfunction
04/12/81666500
111/12/81670501
23/22/82669500
311/11/82668500
44/04/83667500
56/18/82672500
68/30/836731000
711/28/836701000
82/03/846572001
94/06/846632001
108/30/846702001
1110/05/846782000
1211/08/846672000
131/24/856532002
144/12/856672000
154/29/856752000
166/17/856702000
177/2903/856812000
188/27/856762000
1910/03/856792000
2010/30/856752002
2111/26/856762000
221/12/866582001
\n", "
" ], "text/plain": [ " Date Count Temperature Pressure Malfunction\n", "0 4/12/81 6 66 50 0\n", "1 11/12/81 6 70 50 1\n", "2 3/22/82 6 69 50 0\n", "3 11/11/82 6 68 50 0\n", "4 4/04/83 6 67 50 0\n", "5 6/18/82 6 72 50 0\n", "6 8/30/83 6 73 100 0\n", "7 11/28/83 6 70 100 0\n", "8 2/03/84 6 57 200 1\n", "9 4/06/84 6 63 200 1\n", "10 8/30/84 6 70 200 1\n", "11 10/05/84 6 78 200 0\n", "12 11/08/84 6 67 200 0\n", "13 1/24/85 6 53 200 2\n", "14 4/12/85 6 67 200 0\n", "15 4/29/85 6 75 200 0\n", "16 6/17/85 6 70 200 0\n", "17 7/2903/85 6 81 200 0\n", "18 8/27/85 6 76 200 0\n", "19 10/03/85 6 79 200 0\n", "20 10/30/85 6 75 200 2\n", "21 11/26/85 6 76 200 0\n", "22 1/12/86 6 58 200 1" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "#data = pd.read_csv(\"https://app-learninglab.inria.fr/moocrr/gitlab/moocrr-session3/moocrr-reproducibility-study/blob/master/data/shuttle.csv\")\n", "data_url = 'https://app-learninglab.inria.fr/moocrr/gitlab/moocrr-session3/moocrr-reproducibility-study/blob/master/data/shuttle.csv'\n", "\n", "data_file='/home/aschmide/Documents/formation_MOOC_RR/mooc-rr/module4/shuttle.csv'\n", "import os\n", "import urllib.request\n", "\n", "if not os.path.exists(data_file):\n", " urllib.request.urlretrieve(data_url, data_file)\n", " \n", "data = pd.read_csv(data_file)\n", "\n", "import time\n", "modtime=os.path.getmtime(data_file)\n", "modtime=time.ctime(modtime)\n", "print('file downloaded on',modtime)\n", "data" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We know from our previous experience on this data set that filtering data is a really bad idea. We will therefore process it as such." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "%matplotlib inline\n", "pd.set_option('mode.chained_assignment',None) # this removes a useless warning from pandas\n", "import matplotlib.pyplot as plt\n", "\n", "data[\"Frequency\"]=data.Malfunction/data.Count\n", "data.plot(x=\"Temperature\",y=\"Frequency\",kind=\"scatter\",ylim=[0,1])\n", "plt.grid(True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Logistic regression\n", "\n", "Let's assume O-rings independently fail with the same probability which solely depends on temperature. A logistic regression should allow us to estimate the influence of temperature." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "import statsmodels.api as sm\n", "\n", "data[\"Success\"]=data.Count-data.Malfunction\n", "data[\"Intercept\"]=1\n", "\n", "logmodel=sm.GLM(data['Frequency'], data[['Intercept','Temperature']], \n", " family=sm.families.Binomial(sm.families.links.logit)).fit()\n", "\n", "logmodel.summary()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The maximum likelyhood estimator of the intercept and of Temperature are thus $\\hat{\\alpha}=5.0849$ and $\\hat{\\beta}=-0.1156$. This **corresponds** to the values from the article of Dalal *et al.* The standard errors are $s_{\\hat{\\alpha}} = 7.477$ and $s_{\\hat{\\beta}} = 0.115$, which is **different** from the $3.052$ and $0.04702$ reported by Dallal *et al.* The deviance is $3.01444$ with 21 degrees of freedom. I cannot find any value similar to the Goodness of fit ($G^2=18.086$) reported by Dalal *et al.* There seems to be something wrong. Oh I know, I haven't indicated that my observations are actually the result of 6 observations for each rocket launch. Let's indicate these weights (since the weights are always the same throughout all experiments, it does not change the estimates of the fit but it does influence the variance estimates)." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "logmodel=sm.GLM(data['Frequency'], data[['Intercept','Temperature']], \n", " family=sm.families.Binomial(sm.families.links.logit),\n", " var_weights=data['Count']).fit()\n", "\n", "logmodel.summary()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Good, now I have recovered the asymptotic standard errors $s_{\\hat{\\alpha}}=3.052$ and $s_{\\hat{\\beta}}=0.047$.\n", "The Goodness of fit (Deviance) indicated for this model is $G^2=18.086$ with 21 degrees of freedom (Df Residuals).\n", "\n", "**I have therefore managed to fully replicate the results of the Dalal *et al.* article**." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Predicting failure probability\n", "The temperature when launching the shuttle was 31°F. Let's try to estimate the failure probability for such temperature using our model.:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "%matplotlib inline\n", "data_pred = pd.DataFrame({'Temperature': np.linspace(start=30, stop=90, num=121), 'Intercept': 1})\n", "data_pred['Frequency'] = logmodel.predict(data_pred)\n", "data_pred.plot(x=\"Temperature\",y=\"Frequency\",kind=\"line\",ylim=[0,1])\n", "plt.scatter(x=data[\"Temperature\"],y=data[\"Frequency\"])\n", "plt.grid(True)" ] }, { "cell_type": "markdown", "metadata": { "hideCode": false, "hidePrompt": false, "scrolled": true }, "source": [ "This figure is very similar to the Figure 4 of Dalal *et al.* **I have managed to replicate the Figure 4 of the Dalal *et al.* article.**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Computing and plotting uncertainty" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Following the documentation of [Seaborn](https://seaborn.pydata.org/generated/seaborn.regplot.html), I use regplot." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "sns.set(color_codes=True)\n", "plt.xlim(30,90)\n", "plt.ylim(0,1)\n", "sns.regplot(x='Temperature', y='Frequency', data=data, logistic=True)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**I think I have managed to correctly compute and plot the uncertainty of my prediction.** Although the shaded area seems very similar to [the one obtained by with R](https://app-learninglab.inria.fr/moocrr/gitlab/moocrr-session3/moocrr-reproducibility-study/tree/master/challenger.pdf), I can spot a few differences (e.g., the blue point for temperature 63 is outside)... Could this be a numerical error ? Or a difference in the statistical method ? It is not clear which one is \"right\"." ] } ], "metadata": { "celltoolbar": "Hide code", "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": "3.7.6" }, "toc": { "base_numbering": 1, "nav_menu": {}, "number_sections": true, "sideBar": true, "skip_h1_title": false, "title_cell": "Table of Contents", "title_sidebar": "Contents", "toc_cell": false, "toc_position": {}, "toc_section_display": true, "toc_window_display": false } }, "nbformat": 4, "nbformat_minor": 2 }