{"cells": [{"cell_type": "markdown", "metadata": {}, "source": ["# 2A.ml - Statistiques descriptives avec scikit-learn\n", "\n", "[ACP](https://fr.wikipedia.org/wiki/Analyse_en_composantes_principales), [CAH](https://en.wikipedia.org/wiki/Hierarchical_clustering), [r\u00e9gression lineaire](https://fr.wikipedia.org/wiki/R%C3%A9gression_lin%C3%A9aire)."]}, {"cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": ["%matplotlib inline"]}, {"cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": ["import matplotlib.pyplot as plt"]}, {"cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [{"data": {"text/html": ["<div id=\"my_id_menu_nb\">run previous cell, wait for 2 seconds</div>\n", "<script>\n", "function repeat_indent_string(n){\n", "    var a = \"\" ;\n", "    for ( ; n > 0 ; --n)\n", "        a += \"    \";\n", "    return a;\n", "}\n", "// look up into all sections and builds an automated menu //\n", "var update_menu_string = function(begin, lfirst, llast, sformat, send, keep_item, begin_format, end_format) {\n", "    var anchors = document.getElementsByClassName(\"section\");\n", "    if (anchors.length == 0) {\n", "        anchors = document.getElementsByClassName(\"text_cell_render rendered_html\");\n", "    }\n", "    var i,t;\n", "    var text_menu = begin;\n", "    var text_memo = \"<pre>\\nlength:\" + anchors.length + \"\\n\";\n", "    var ind = \"\";\n", "    var memo_level = 1;\n", "    var href;\n", "    var tags = [];\n", "    var main_item = 0;\n", "    var format_open = 0;\n", "    for (i = 0; i <= llast; i++)\n", "        tags.push(\"h\" + i);\n", "\n", "    for (i = 0; i < anchors.length; i++) {\n", "        text_memo += \"**\" + anchors[i].id + \"--\\n\";\n", "\n", "        var child = null;\n", "        for(t = 0; t < tags.length; t++) {\n", "            var r = anchors[i].getElementsByTagName(tags[t]);\n", "            if (r.length > 0) {\n", "child = r[0];\n", "break;\n", "            }\n", "        }\n", "        if (child == null) {\n", "            text_memo += \"null\\n\";\n", "            continue;\n", "        }\n", "        if (anchors[i].hasAttribute(\"id\")) {\n", "            // when converted in RST\n", "            href = anchors[i].id;\n", "            text_memo += \"#1-\" + href;\n", "            // passer \u00e0 child suivant (le chercher)\n", "        }\n", "        else if (child.hasAttribute(\"id\")) {\n", "            // in a notebook\n", "            href = child.id;\n", "            text_memo += \"#2-\" + href;\n", "        }\n", "        else {\n", "            text_memo += \"#3-\" + \"*\" + \"\\n\";\n", "            continue;\n", "        }\n", "        var title = child.textContent;\n", "        var level = parseInt(child.tagName.substring(1,2));\n", "\n", "        text_memo += \"--\" + level + \"?\" + lfirst + \"--\" + title + \"\\n\";\n", "\n", "        if ((level < lfirst) || (level > llast)) {\n", "            continue ;\n", "        }\n", "        if (title.endsWith('\u00b6')) {\n", "            title = title.substring(0,title.length-1).replace(\"<\", \"&lt;\")\n", "         .replace(\">\", \"&gt;\").replace(\"&\", \"&amp;\");\n", "        }\n", "        if (title.length == 0) {\n", "            continue;\n", "        }\n", "\n", "        while (level < memo_level) {\n", "            text_menu += end_format + \"</ul>\\n\";\n", "            format_open -= 1;\n", "            memo_level -= 1;\n", "        }\n", "        if (level == lfirst) {\n", "            main_item += 1;\n", "        }\n", "        if (keep_item != -1 && main_item != keep_item + 1) {\n", "            // alert(main_item + \" - \" + level + \" - \" + keep_item);\n", "            continue;\n", "        }\n", "        while (level > memo_level) {\n", "            text_menu += \"<ul>\\n\";\n", "            memo_level += 1;\n", "        }\n", "        text_menu += repeat_indent_string(level-2);\n", "        text_menu += begin_format + sformat.replace(\"__HREF__\", href).replace(\"__TITLE__\", title);\n", "        format_open += 1;\n", "    }\n", "    while (1 < memo_level) {\n", "        text_menu += end_format + \"</ul>\\n\";\n", "        memo_level -= 1;\n", "        format_open -= 1;\n", "    }\n", "    text_menu += send;\n", "    //text_menu += \"\\n\" + text_memo;\n", "\n", "    while (format_open > 0) {\n", "        text_menu += end_format;\n", "        format_open -= 1;\n", "    }\n", "    return text_menu;\n", "};\n", "var update_menu = function() {\n", "    var sbegin = \"\";\n", "    var sformat = '<a href=\"#__HREF__\">__TITLE__</a>';\n", "    var send = \"\";\n", "    var begin_format = '<li>';\n", "    var end_format = '</li>';\n", "    var keep_item = -1;\n", "    var text_menu = update_menu_string(sbegin, 2, 4, sformat, send, keep_item,\n", "       begin_format, end_format);\n", "    var menu = document.getElementById(\"my_id_menu_nb\");\n", "    menu.innerHTML=text_menu;\n", "};\n", "window.setTimeout(update_menu,2000);\n", "            </script>"], "text/plain": ["<IPython.core.display.HTML object>"]}, "execution_count": 4, "metadata": {}, "output_type": "execute_result"}], "source": ["from jyquickhelper import add_notebook_menu\n", "add_notebook_menu()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Introduction\n", "\n", "Les statistiques descriptives sont abord\u00e9es de la premi\u00e8re ann\u00e9e \u00e0 l'ENSAE. Un des livres que je consulte souvent est celui de Gilles Saporta : [Probabilit\u00e9s, analyse des donn\u00e9es et statistique](http://www.editionstechnip.com/fr/catalogue-detail/149/probabilites-analyse-des-donnees-et-statistique.html) qui est en fran\u00e7ais.\n", "\n", "Le module [scikit-learn](http://scikit-learn.org/stable/) a largement contribu\u00e9 au succ\u00e8s de Python dans le domaine du [machine learning](http://en.wikipedia.org/wiki/Machine_learning). Ce module inclut de nombreuses techniques regroup\u00e9es sous le terme statistiques descriptives. La correspondance anglais-fran\u00e7ais n'est pas toujours \u00e9vidente. Voici quelques termes :\n", "\n", "* [ACP](http://fr.wikipedia.org/wiki/ACP) - [PCA](http://scikit-learn.org/stable/modules/decomposition.html#decompositions)\n", "* [k-moyennes](http://fr.wikipedia.org/wiki/Algorithme_des_k-moyennes) - [k-means](http://scikit-learn.org/stable/modules/clustering.html#k-means)\n", "* [CAH](http://en.wikipedia.org/wiki/CAH) - [Hierarchical Clustering](http://scikit-learn.org/stable/modules/clustering.html#hierarchical-clustering)\n", "* [k plus proches voisins](http://fr.wikipedia.org/wiki/Recherche_des_plus_proches_voisins) - [k-PPV](http://scikit-learn.org/stable/modules/neighbors.html)\n", "* [analyse lin\u00e9aire discriminante](http://fr.wikipedia.org/wiki/Analyse_discriminante_lin%C3%A9aire) - [LDA](http://scikit-learn.org/stable/modules/classes.html#module-sklearn.lda)\n", "* [r\u00e9gression lin\u00e9aire](http://fr.wikipedia.org/wiki/R%C3%A9gression_lin%C3%A9aire) - [linear regression](http://scikit-learn.org/stable/modules/linear_model.html#ordinary-least-squares)\n", "\n", "\n", "[scikit-learn](http://scikit-learn.org/stable/) est orient\u00e9 machine learning, les r\u00e9sultats qu'il produit sont un peu moins complets que [statsmodels](http://statsmodels.sourceforge.net/) pour les mod\u00e8les statistiques lin\u00e9aires ou [fastcluster](http://cran.r-project.org/web/packages/fastcluster/vignettes/fastcluster.pdf) pour la CAH. L'objectif de ces deux heures est d'utiliser ces modules pour \u00e9tudier un jeu de donn\u00e9es :"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### ACP (Analyse en Composantes Principales)\n", "\n", "Le site [data.gouv.fr](https://www.data.gouv.fr/) propose de nombreux jeux de donn\u00e9es dont [S\u00e9ries chronologiques Education : les \u00e9l\u00e8ves du second degr\u00e9](https://www.data.gouv.fr/fr/datasets/series-chronologiques-education-les-eleves-du-second-degre/). Ces donn\u00e9es sont \u00e9galement accessibles comme ceci :"]}, {"cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["(27, 21)\n"]}, {"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>1993</th>\n", "      <th>1994</th>\n", "      <th>1995</th>\n", "      <th>1996</th>\n", "      <th>1997</th>\n", "      <th>1998</th>\n", "      <th>1999</th>\n", "      <th>2000</th>\n", "      <th>2001</th>\n", "      <th>2002</th>\n", "      <th>...</th>\n", "      <th>2004</th>\n", "      <th>2005</th>\n", "      <th>2006</th>\n", "      <th>2007</th>\n", "      <th>2008</th>\n", "      <th>2009</th>\n", "      <th>2010</th>\n", "      <th>2011</th>\n", "      <th>2012</th>\n", "      <th>2013</th>\n", "    </tr>\n", "    <tr>\n", "      <th>acad\u00e9mie</th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>Aix-Marseille</th>\n", "      <td>241357</td>\n", "      <td>242298</td>\n", "      <td>242096</td>\n", "      <td>242295</td>\n", "      <td>243660</td>\n", "      <td>244608</td>\n", "      <td>245536</td>\n", "      <td>247288</td>\n", "      <td>249331</td>\n", "      <td>250871</td>\n", "      <td>...</td>\n", "      <td>250622</td>\n", "      <td>248208</td>\n", "      <td>245755</td>\n", "      <td>243832</td>\n", "      <td>242309</td>\n", "      <td>240664</td>\n", "      <td>240432</td>\n", "      <td>241336</td>\n", "      <td>239051</td>\n", "      <td>240115</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Amiens</th>\n", "      <td>198281</td>\n", "      <td>196871</td>\n", "      <td>195709</td>\n", "      <td>194055</td>\n", "      <td>192893</td>\n", "      <td>191862</td>\n", "      <td>189636</td>\n", "      <td>185977</td>\n", "      <td>183357</td>\n", "      <td>180973</td>\n", "      <td>...</td>\n", "      <td>175610</td>\n", "      <td>172110</td>\n", "      <td>168718</td>\n", "      <td>165295</td>\n", "      <td>163116</td>\n", "      <td>162548</td>\n", "      <td>163270</td>\n", "      <td>164422</td>\n", "      <td>165275</td>\n", "      <td>166345</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Besan\u00e7on</th>\n", "      <td>116373</td>\n", "      <td>115600</td>\n", "      <td>114282</td>\n", "      <td>113312</td>\n", "      <td>112076</td>\n", "      <td>110261</td>\n", "      <td>108106</td>\n", "      <td>105463</td>\n", "      <td>103336</td>\n", "      <td>102264</td>\n", "      <td>...</td>\n", "      <td>100117</td>\n", "      <td>98611</td>\n", "      <td>97038</td>\n", "      <td>95779</td>\n", "      <td>95074</td>\n", "      <td>94501</td>\n", "      <td>94599</td>\n", "      <td>94745</td>\n", "      <td>94351</td>\n", "      <td>94613</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Bordeaux</th>\n", "      <td>253551</td>\n", "      <td>252644</td>\n", "      <td>249658</td>\n", "      <td>247708</td>\n", "      <td>247499</td>\n", "      <td>245757</td>\n", "      <td>244992</td>\n", "      <td>243047</td>\n", "      <td>243592</td>\n", "      <td>245198</td>\n", "      <td>...</td>\n", "      <td>244805</td>\n", "      <td>244343</td>\n", "      <td>242602</td>\n", "      <td>242933</td>\n", "      <td>243146</td>\n", "      <td>244336</td>\n", "      <td>246806</td>\n", "      <td>250626</td>\n", "      <td>252085</td>\n", "      <td>255761</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Caen</th>\n", "      <td>145435</td>\n", "      <td>144369</td>\n", "      <td>141883</td>\n", "      <td>140658</td>\n", "      <td>139585</td>\n", "      <td>137704</td>\n", "      <td>135613</td>\n", "      <td>133255</td>\n", "      <td>131206</td>\n", "      <td>129271</td>\n", "      <td>...</td>\n", "      <td>125552</td>\n", "      <td>123889</td>\n", "      <td>122550</td>\n", "      <td>121002</td>\n", "      <td>119857</td>\n", "      <td>119426</td>\n", "      <td>119184</td>\n", "      <td>119764</td>\n", "      <td>119010</td>\n", "      <td>119238</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "<p>5 rows \u00d7 21 columns</p>\n", "</div>"], "text/plain": ["                 1993    1994    1995    1996    1997    1998    1999    2000  \\\n", "acad\u00e9mie                                                                        \n", "Aix-Marseille  241357  242298  242096  242295  243660  244608  245536  247288   \n", "Amiens         198281  196871  195709  194055  192893  191862  189636  185977   \n", "Besan\u00e7on       116373  115600  114282  113312  112076  110261  108106  105463   \n", "Bordeaux       253551  252644  249658  247708  247499  245757  244992  243047   \n", "Caen           145435  144369  141883  140658  139585  137704  135613  133255   \n", "\n", "                 2001    2002  ...    2004    2005    2006    2007    2008  \\\n", "acad\u00e9mie                       ...                                           \n", "Aix-Marseille  249331  250871  ...  250622  248208  245755  243832  242309   \n", "Amiens         183357  180973  ...  175610  172110  168718  165295  163116   \n", "Besan\u00e7on       103336  102264  ...  100117   98611   97038   95779   95074   \n", "Bordeaux       243592  245198  ...  244805  244343  242602  242933  243146   \n", "Caen           131206  129271  ...  125552  123889  122550  121002  119857   \n", "\n", "                 2009    2010    2011    2012    2013  \n", "acad\u00e9mie                                               \n", "Aix-Marseille  240664  240432  241336  239051  240115  \n", "Amiens         162548  163270  164422  165275  166345  \n", "Besan\u00e7on        94501   94599   94745   94351   94613  \n", "Bordeaux       244336  246806  250626  252085  255761  \n", "Caen           119426  119184  119764  119010  119238  \n", "\n", "[5 rows x 21 columns]"]}, "execution_count": 5, "metadata": {}, "output_type": "execute_result"}], "source": ["import pandas, numpy, pyensae.datasource\n", "from sklearn.decomposition import PCA\n", "import matplotlib.pyplot as plt\n", "\n", "fichier = pyensae.datasource.download_data(\"eleve_region.txt\")\n", "df = pandas.read_csv(\"eleve_region.txt\", sep=\"\\t\", encoding=\"utf8\", index_col=0)\n", "print(df.shape)\n", "df.head(n=5)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On veut observer l'\u00e9volution du nombre d'\u00e9l\u00e8ves. On prend l'ann\u00e9e 1993 comme base 100."]}, {"cell_type": "code", "execution_count": 5, "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>1993</th>\n", "      <th>1994</th>\n", "      <th>1995</th>\n", "      <th>1996</th>\n", "      <th>1997</th>\n", "      <th>1998</th>\n", "      <th>1999</th>\n", "      <th>2000</th>\n", "      <th>2001</th>\n", "      <th>2002</th>\n", "      <th>...</th>\n", "      <th>2004</th>\n", "      <th>2005</th>\n", "      <th>2006</th>\n", "      <th>2007</th>\n", "      <th>2008</th>\n", "      <th>2009</th>\n", "      <th>2010</th>\n", "      <th>2011</th>\n", "      <th>2012</th>\n", "      <th>2013</th>\n", "    </tr>\n", "    <tr>\n", "      <th>acad\u00e9mie</th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "      <th></th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>Aix-Marseille</th>\n", "      <td>1.0</td>\n", "      <td>1.003899</td>\n", "      <td>1.003062</td>\n", "      <td>1.003886</td>\n", "      <td>1.009542</td>\n", "      <td>1.013470</td>\n", "      <td>1.017315</td>\n", "      <td>1.024574</td>\n", "      <td>1.033038</td>\n", "      <td>1.039419</td>\n", "      <td>...</td>\n", "      <td>1.038387</td>\n", "      <td>1.028385</td>\n", "      <td>1.018222</td>\n", "      <td>1.010255</td>\n", "      <td>1.003944</td>\n", "      <td>0.997129</td>\n", "      <td>0.996168</td>\n", "      <td>0.999913</td>\n", "      <td>0.990446</td>\n", "      <td>0.994854</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Amiens</th>\n", "      <td>1.0</td>\n", "      <td>0.992889</td>\n", "      <td>0.987029</td>\n", "      <td>0.978687</td>\n", "      <td>0.972826</td>\n", "      <td>0.967627</td>\n", "      <td>0.956400</td>\n", "      <td>0.937947</td>\n", "      <td>0.924733</td>\n", "      <td>0.912710</td>\n", "      <td>...</td>\n", "      <td>0.885662</td>\n", "      <td>0.868011</td>\n", "      <td>0.850904</td>\n", "      <td>0.833640</td>\n", "      <td>0.822651</td>\n", "      <td>0.819786</td>\n", "      <td>0.823427</td>\n", "      <td>0.829237</td>\n", "      <td>0.833539</td>\n", "      <td>0.838936</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Besan\u00e7on</th>\n", "      <td>1.0</td>\n", "      <td>0.993358</td>\n", "      <td>0.982032</td>\n", "      <td>0.973697</td>\n", "      <td>0.963076</td>\n", "      <td>0.947479</td>\n", "      <td>0.928961</td>\n", "      <td>0.906250</td>\n", "      <td>0.887972</td>\n", "      <td>0.878761</td>\n", "      <td>...</td>\n", "      <td>0.860311</td>\n", "      <td>0.847370</td>\n", "      <td>0.833853</td>\n", "      <td>0.823035</td>\n", "      <td>0.816976</td>\n", "      <td>0.812053</td>\n", "      <td>0.812895</td>\n", "      <td>0.814149</td>\n", "      <td>0.810764</td>\n", "      <td>0.813015</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Bordeaux</th>\n", "      <td>1.0</td>\n", "      <td>0.996423</td>\n", "      <td>0.984646</td>\n", "      <td>0.976955</td>\n", "      <td>0.976131</td>\n", "      <td>0.969261</td>\n", "      <td>0.966243</td>\n", "      <td>0.958572</td>\n", "      <td>0.960722</td>\n", "      <td>0.967056</td>\n", "      <td>...</td>\n", "      <td>0.965506</td>\n", "      <td>0.963684</td>\n", "      <td>0.956817</td>\n", "      <td>0.958123</td>\n", "      <td>0.958963</td>\n", "      <td>0.963656</td>\n", "      <td>0.973398</td>\n", "      <td>0.988464</td>\n", "      <td>0.994218</td>\n", "      <td>1.008716</td>\n", "    </tr>\n", "    <tr>\n", "      <th>Caen</th>\n", "      <td>1.0</td>\n", "      <td>0.992670</td>\n", "      <td>0.975577</td>\n", "      <td>0.967154</td>\n", "      <td>0.959776</td>\n", "      <td>0.946842</td>\n", "      <td>0.932465</td>\n", "      <td>0.916251</td>\n", "      <td>0.902162</td>\n", "      <td>0.888858</td>\n", "      <td>...</td>\n", "      <td>0.863286</td>\n", "      <td>0.851851</td>\n", "      <td>0.842644</td>\n", "      <td>0.832001</td>\n", "      <td>0.824128</td>\n", "      <td>0.821164</td>\n", "      <td>0.819500</td>\n", "      <td>0.823488</td>\n", "      <td>0.818304</td>\n", "      <td>0.819871</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "<p>5 rows \u00d7 21 columns</p>\n", "</div>"], "text/plain": ["               1993      1994      1995      1996      1997      1998  \\\n", "acad\u00e9mie                                                                \n", "Aix-Marseille   1.0  1.003899  1.003062  1.003886  1.009542  1.013470   \n", "Amiens          1.0  0.992889  0.987029  0.978687  0.972826  0.967627   \n", "Besan\u00e7on        1.0  0.993358  0.982032  0.973697  0.963076  0.947479   \n", "Bordeaux        1.0  0.996423  0.984646  0.976955  0.976131  0.969261   \n", "Caen            1.0  0.992670  0.975577  0.967154  0.959776  0.946842   \n", "\n", "                   1999      2000      2001      2002  ...      2004  \\\n", "acad\u00e9mie                                               ...             \n", "Aix-Marseille  1.017315  1.024574  1.033038  1.039419  ...  1.038387   \n", "Amiens         0.956400  0.937947  0.924733  0.912710  ...  0.885662   \n", "Besan\u00e7on       0.928961  0.906250  0.887972  0.878761  ...  0.860311   \n", "Bordeaux       0.966243  0.958572  0.960722  0.967056  ...  0.965506   \n", "Caen           0.932465  0.916251  0.902162  0.888858  ...  0.863286   \n", "\n", "                   2005      2006      2007      2008      2009      2010  \\\n", "acad\u00e9mie                                                                    \n", "Aix-Marseille  1.028385  1.018222  1.010255  1.003944  0.997129  0.996168   \n", "Amiens         0.868011  0.850904  0.833640  0.822651  0.819786  0.823427   \n", "Besan\u00e7on       0.847370  0.833853  0.823035  0.816976  0.812053  0.812895   \n", "Bordeaux       0.963684  0.956817  0.958123  0.958963  0.963656  0.973398   \n", "Caen           0.851851  0.842644  0.832001  0.824128  0.821164  0.819500   \n", "\n", "                   2011      2012      2013  \n", "acad\u00e9mie                                     \n", "Aix-Marseille  0.999913  0.990446  0.994854  \n", "Amiens         0.829237  0.833539  0.838936  \n", "Besan\u00e7on       0.814149  0.810764  0.813015  \n", "Bordeaux       0.988464  0.994218  1.008716  \n", "Caen           0.823488  0.818304  0.819871  \n", "\n", "[5 rows x 21 columns]"]}, "execution_count": 6, "metadata": {}, "output_type": "execute_result"}], "source": ["for c in df.columns:\n", "    if c != \"1993\":\n", "        df[c] /= df [\"1993\"]\n", "df[\"1993\"] /= df[\"1993\"]\n", "df.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Il n'est pas \u00e9vident d'analyser ce tableaux de chiffres. On utilise une ACP pour projeter les acad\u00e9mies dans un plan."]}, {"cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["PCA(n_components=4)\n"]}, {"data": {"text/plain": ["array([9.63024476e-01, 3.37267932e-02, 1.93116666e-03, 6.61046399e-04])"]}, "execution_count": 7, "metadata": {}, "output_type": "execute_result"}], "source": ["pca = PCA(n_components=4)\n", "print(pca.fit(df))\n", "pca.explained_variance_ratio_"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Le premier axe explique l'essentiel de la variance. Les variables n'ont pas \u00e9t\u00e9 normalis\u00e9es car elles \u00e9voluent dans les m\u00eames ordres de grandeur."]}, {"cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 432x288 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["plt.bar(numpy.arange(len(pca.explained_variance_ratio_)) + 0.5, \n", "        pca.explained_variance_ratio_)\n", "plt.title(\"Variance expliqu\u00e9e\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On affiche les acad\u00e9mies dans le plan des deux premiers axes :"]}, {"cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 1296x432 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["X_reduced = pca.transform(df)\n", "plt.figure(figsize=(18, 6))\n", "plt.scatter(X_reduced[:, 0], X_reduced[:, 1])\n", "\n", "for label, x, y in zip(df.index, X_reduced[:, 0], X_reduced[:, 1]):\n", "    plt.annotate(\n", "        label, \n", "        xy=(x, y), xytext=(-10, 10),\n", "        textcoords='offset points', ha='right', va='bottom',\n", "        bbox=dict(boxstyle='round,pad=0.5', fc='yellow', alpha=0.5),\n", "        arrowprops = dict(arrowstyle='->', connectionstyle='arc3,rad=0'))\n", "plt.title(\"ACP rapprochant des profils d'\u00e9volution similaires\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Puis on v\u00e9rifie que deux villes proches ont le m\u00eame profil d'\u00e9volution au cours des ann\u00e9es :"]}, {"cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 720x288 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["sub = df.loc[[\"Paris\", \"Bordeaux\", \"Lyon\", \"Nice\", \"La R\u00e9union\", \"Reims\"], :]\n", "ax = sub.transpose().plot(figsize=(10, 4))\n", "ax.set_title(\"Evolution des effectifs au cours du temps\");"]}, {"cell_type": "markdown", "metadata": {}, "source": ["L'[ACP version statsmodels](http://statsmodels.sourceforge.net/devel/generated/statsmodels.sandbox.tools.tools_pca.pca.html) produit le m\u00eame type de r\u00e9sultats. Un exemple est disponible ici : [PCA and Biplot using Python](http://okomestudio.net/biboroku/?p=2292)."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Exercice 1 : CAH (classification ascendante hi\u00e9rarchique)\n", "\n", "Le point commun de ces m\u00e9thodes est qu'elles ne sont pas supervis\u00e9es. L'objectif est de r\u00e9duire la complexit\u00e9 des donn\u00e9es. R\u00e9duire le nombre de dimensions pour l'ACP ou segmenter les observations pour les k-means et la CAH. On propose d'utiliser une CAH sur les m\u00eames donn\u00e9es.\n", "\n", "Le module [scikit-learn.cluster](http://scikit-learn.org/stable/modules/clustering.html#hierarchical-clustering) ne propose pas de fonction pour dessiner le [dendrogram](http://en.wikipedia.org/wiki/Dendrogram). Il faudra utiliser celle-ci : [dendrogram](http://docs.scipy.org/doc/scipy/reference/generated/scipy.cluster.hierarchy.dendrogram.html) et sans doute s'inspirer du code suivant."]}, {"cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": ["<Figure size 576x576 with 1 Axes>"]}, "metadata": {"needs_background": "light"}, "output_type": "display_data"}], "source": ["from sklearn.cluster import AgglomerativeClustering\n", "from scipy.cluster.hierarchy import dendrogram\n", "\n", "ward = AgglomerativeClustering(linkage='ward', compute_full_tree=True).fit(df)\n", "dendro = [ ]\n", "for a,b in ward.children_:\n", "    dendro.append([a, b, float(len(dendro)+1), len(dendro)+1])\n", "    \n", "fig = plt.figure(figsize=(8, 8))\n", "ax = fig.add_subplot(1, 1, 1) \n", "r = dendrogram(dendro, color_threshold=1, labels=list(df.index),\n", "               show_leaf_counts=True, ax=ax, orientation=\"left\")    "]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Exercice 2 : r\u00e9gression\n", "\n", "Ce sont trois m\u00e9thodes supervis\u00e9es : on s'en sert pour expliquer pr\u00e9dire le lien entre deux variables $X$ et $Y$ (ou ensemble de variables) ou pr\u00e9dire $Y$ en fonction de $X$. Pour cet exercice, on r\u00e9cup\u00e8re des donn\u00e9es relatives aux salaires [Salaires et revenus d'activit\u00e9s](https://www.insee.fr/fr/statistiques/2011542?sommaire=2011795) (les chercher avec la requ\u00eate *insee donn\u00e9es dads* sur un moteur de recherche). La r\u00e9cup\u00e9ration des donn\u00e9es est assez fastidieuse. La premi\u00e8re \u00e9tape consiste \u00e0 t\u00e9l\u00e9charger les donn\u00e9es depuis le site de l'[INSEE](http://www.insee.fr/). La seconde \u00e9tape consiste \u00e0 convertir les donn\u00e9es au format [sqlite3](https://docs.python.org/3.4/library/sqlite3.html). Pour ce fichier, ce travail a d\u00e9j\u00e0 \u00e9t\u00e9 effectu\u00e9 et peut \u00eatre t\u00e9l\u00e9charg\u00e9 depuis mon site. La base comprend 2 millions de lignes."]}, {"cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [{"data": {"text/plain": ["['.\\\\salaries11.dbf', 'varlist_salaries11.dbf', '.\\\\varmod_salaries11.dbf']"]}, "execution_count": 12, "metadata": {}, "output_type": "execute_result"}], "source": ["import pyensae.datasource\n", "f = pyensae.datasource.download_data(\"dads2011_gf_salaries11_dbase.zip\",\n", "                          website=\"https://www.insee.fr/fr/statistiques/fichier/2011542/\")\n", "f"]}, {"cell_type": "code", "execution_count": 12, "metadata": {"scrolled": false}, "outputs": [{"name": "stdout", "output_type": "stream", "text": ["convert dbase into sqllite\n", "SQL  'drop table if exists salaries11'\n", "SQL  'create table salaries (A6 TEXT, A17 TEXT, A38 TEXT, REGR TEXT, DEPR TEXT, REGT TEXT, DEPT TEXT, SEXE TEXT, PCS TEXT, CONT_TRAV TEXT, CONV_COLL TEXT, TYP_EMPLOI TEXT, DUREE REAL, DATDEB REAL, DATFIN REAL, CPFD TEXT, DOMEMPL TEXT, DOMEMPL_EM TEXT, FILT TEXT, AGE REAL, CS TEXT, NB_PER REAL, NB_PER_N REAL, NBHEUR REAL, NBHEUR_TOT REAL, TRALCHT TEXT, TREFF TEXT, TRNNETO TEXT, POND REAL)'\n", "moving line  0  to table salaries\n", "moving line  20000  to table salaries\n", "moving line  40000  to table salaries\n", "moving line  60000  to table salaries\n", "moving line  80000  to table salaries\n", "moving line  100000  to table salaries\n", "moving line  120000  to table salaries\n", "moving line  140000  to table salaries\n", "moving line  160000  to table salaries\n", "moving line  180000  to table salaries\n", "moving line  200000  to table salaries\n", "moving line  220000  to table salaries\n", "moving line  240000  to table salaries\n", "moving line  260000  to table salaries\n", "moving line  280000  to table salaries\n", "moving line  300000  to table salaries\n", "moving line  320000  to table salaries\n", "moving line  340000  to table salaries\n", "moving line  360000  to table salaries\n", "moving line  380000  to table salaries\n", "moving line  400000  to table salaries\n", "moving line  420000  to table salaries\n", "moving line  440000  to table salaries\n", "moving line  460000  to table salaries\n", "moving line  480000  to table salaries\n", "moving line  500000  to table salaries\n", "moving line  520000  to table salaries\n", "moving line  540000  to table salaries\n", "moving line  560000  to table salaries\n", "moving line  580000  to table salaries\n", "moving line  600000  to table salaries\n", "moving line  620000  to table salaries\n", "moving line  640000  to table salaries\n", "moving line  660000  to table salaries\n", "moving line  680000  to table salaries\n", "moving line  700000  to table salaries\n", "moving line  720000  to table salaries\n", "moving line  740000  to table salaries\n", "moving line  760000  to table salaries\n", "moving line  780000  to table salaries\n", "moving line  800000  to table salaries\n", "moving line  820000  to table salaries\n", "moving line  840000  to table salaries\n", "moving line  860000  to table salaries\n", "moving line  880000  to table salaries\n", "moving line  900000  to table salaries\n", "moving line  920000  to table salaries\n", "moving line  940000  to table salaries\n", "moving line  960000  to table salaries\n", "moving line  980000  to table salaries\n", "moving line  1000000  to table salaries\n", "moving line  1020000  to table salaries\n", "moving line  1040000  to table salaries\n", "moving line  1060000  to table salaries\n", "moving line  1080000  to table salaries\n", "moving line  1100000  to table salaries\n", "moving line  1120000  to table salaries\n", "moving line  1140000  to table salaries\n", "moving line  1160000  to table salaries\n", "moving line  1180000  to table salaries\n", "moving line  1200000  to table salaries\n", "moving line  1220000  to table salaries\n", "moving line  1240000  to table salaries\n", "moving line  1260000  to table salaries\n", "moving line  1280000  to table salaries\n", "moving line  1300000  to table salaries\n", "moving line  1320000  to table salaries\n", "moving line  1340000  to table salaries\n", "moving line  1360000  to table salaries\n", "moving line  1380000  to table salaries\n", "moving line  1400000  to table salaries\n", "moving line  1420000  to table salaries\n", "moving line  1440000  to table salaries\n", "moving line  1460000  to table salaries\n", "moving line  1480000  to table salaries\n", "moving line  1500000  to table salaries\n", "moving line  1520000  to table salaries\n", "moving line  1540000  to table salaries\n", "moving line  1560000  to table salaries\n", "moving line  1580000  to table salaries\n", "moving line  1600000  to table salaries\n", "moving line  1620000  to table salaries\n", "moving line  1640000  to table salaries\n", "moving line  1660000  to table salaries\n", "moving line  1680000  to table salaries\n", "moving line  1700000  to table salaries\n", "moving line  1720000  to table salaries\n", "moving line  1740000  to table salaries\n", "moving line  1760000  to table salaries\n", "moving line  1780000  to table salaries\n", "moving line  1800000  to table salaries\n", "moving line  1820000  to table salaries\n", "moving line  1840000  to table salaries\n", "moving line  1860000  to table salaries\n", "moving line  1880000  to table salaries\n", "moving line  1900000  to table salaries\n", "moving line  1920000  to table salaries\n", "moving line  1940000  to table salaries\n", "moving line  1960000  to table salaries\n", "moving line  1980000  to table salaries\n", "moving line  2000000  to table salaries\n", "moving line  2020000  to table salaries\n", "moving line  2040000  to table salaries\n", "moving line  2060000  to table salaries\n", "moving line  2080000  to table salaries\n", "moving line  2100000  to table salaries\n", "moving line  2120000  to table salaries\n", "moving line  2140000  to table salaries\n", "moving line  2160000  to table salaries\n", "moving line  2180000  to table salaries\n", "moving line  2200000  to table salaries\n", "moving line  2220000  to table salaries\n", "moving line  2240000  to table salaries\n"]}], "source": ["import pandas\n", "try:\n", "    from dbfread import DBF\n", "    use_dbfread = True\n", "except ImportError as e :\n", "    use_dbfread = False\n", "    \n", "if use_dbfread:\n", "    import os\n", "    from pyensae.sql.database_exception import ExceptionSQL\n", "    from pyensae.datasource import dBase2sqllite\n", "    print(\"convert dbase into sqllite\")\n", "    try:\n", "        dBase2sqllite(\"salaries2011.db3\", \"varlist_salaries11.dbf\", overwrite_table=\"varlist\")\n", "        dBase2sqllite(\"salaries2011.db3\", \"varmod_salaries11.dbf\", overwrite_table=\"varmod\")\n", "        dBase2sqllite(\"salaries2011.db3\", 'salaries11.dbf',\n", "                      overwrite_table=\"salaries\", fLOG = print)\n", "    except ExceptionSQL:\n", "        print(\"La base de donn\u00e9es est d\u00e9j\u00e0 renseign\u00e9e.\")\n", "else :\n", "    print(\"use of zipped version\")\n", "    import pyensae.datasource\n", "    db3 = pyensae.datasource.download_data(\"salaries2011.zip\")\n", "    # pour aller plus vite, donn\u00e9es \u00e0 t\u00e9l\u00e9charger au\n", "    # http://www.xavierdupre.fr/enseignement/complements/salaries2011.zip"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Les donn\u00e9es des salaires ne sont pas num\u00e9riques, elles correspondent \u00e0 des intervalles qu'on convertit en prenant le milieu de l'intervalle. Pour le dernier, on prend la borne sup\u00e9rieure."]}, {"cell_type": "code", "execution_count": 13, "metadata": {"scrolled": false}, "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>VARIABLE</th>\n", "      <th>MODALITE</th>\n", "      <th>MODLIBELLE</th>\n", "      <th>montant</th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>8957</th>\n", "      <td>TRNNETO</td>\n", "      <td>00</td>\n", "      <td>[0 ; 200[ euros</td>\n", "      <td>100.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>8958</th>\n", "      <td>TRNNETO</td>\n", "      <td>01</td>\n", "      <td>[200 ; 500[ euros</td>\n", "      <td>350.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>8959</th>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>8960</th>\n", "      <td>TRNNETO</td>\n", "      <td>03</td>\n", "      <td>[1 000 ; 1 500[ euros</td>\n", "      <td>1250.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>8961</th>\n", "      <td>TRNNETO</td>\n", "      <td>04</td>\n", "      <td>[1\u00a0500 ; 2 000[ euros</td>\n", "      <td>1750.0</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["     VARIABLE MODALITE             MODLIBELLE  montant\n", "8957  TRNNETO       00        [0 ; 200[ euros    100.0\n", "8958  TRNNETO       01      [200 ; 500[ euros    350.0\n", "8959  TRNNETO       02    [500 ; 1 000[ euros    750.0\n", "8960  TRNNETO       03  [1 000 ; 1 500[ euros   1250.0\n", "8961  TRNNETO       04  [1\u00a0500 ; 2 000[ euros   1750.0"]}, "execution_count": 14, "metadata": {}, "output_type": "execute_result"}], "source": ["import sqlite3, pandas\n", "con = sqlite3.connect(\"salaries2011.db3\")\n", "df = pandas.io.sql.read_sql(\"select * from varmod\", con)\n", "con.close()\n", "\n", "values = df[ df.VARIABLE == \"TRNNETO\"].copy()\n", "\n", "def process_intervalle(s):\n", "    # [14\u00a0000 ; 16 000[ euros\n", "    acc = \"0123456789;+\"\n", "    s0 = \"\".join(c for c in s if c in acc)\n", "    spl = s0.split(';')\n", "    if len(spl) != 2:\n", "        raise ValueError(\"Unable to process '{0}'\".format(s0))\n", "    try:\n", "        a = float(spl[0])\n", "    except Exception as e:\n", "        raise ValueError(\"Cannot interpret '{0}' - {1}\".format(s, spl))\n", "    b = float(spl[1]) if \"+\" not in spl[1] else None\n", "    if b is None:\n", "        return a\n", "    else:\n", "        return (a+b) / 2.0\n", "\n", "values[\"montant\"] = values.apply(lambda r : process_intervalle(r [\"MODLIBELLE\"]), axis = 1)\n", "values.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On cr\u00e9e la base d'apprentissage :"]}, {"cell_type": "code", "execution_count": 14, "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>TRNNETO</th>\n", "      <th>AGE</th>\n", "      <th>SEXE</th>\n", "      <th>VARIABLE</th>\n", "      <th>MODALITE</th>\n", "      <th>MODLIBELLE</th>\n", "      <th>montant</th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>0</th>\n", "      <td>02</td>\n", "      <td>49.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>1</th>\n", "      <td>02</td>\n", "      <td>27.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>2</th>\n", "      <td>02</td>\n", "      <td>22.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>3</th>\n", "      <td>02</td>\n", "      <td>26.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>4</th>\n", "      <td>02</td>\n", "      <td>29.0</td>\n", "      <td>2</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["  TRNNETO   AGE SEXE VARIABLE MODALITE           MODLIBELLE  montant\n", "0      02  49.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0\n", "1      02  27.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0\n", "2      02  22.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0\n", "3      02  26.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0\n", "4      02  29.0    2  TRNNETO       02  [500 ; 1 000[ euros    750.0"]}, "execution_count": 15, "metadata": {}, "output_type": "execute_result"}], "source": ["import sqlite3, pandas\n", "con = sqlite3.connect(\"salaries2011.db3\")\n", "data = pandas.io.sql.read_sql(\"select TRNNETO,AGE,SEXE from salaries\", con)\n", "con.close()\n", "\n", "salaires = data.merge ( values, left_on=\"TRNNETO\", right_on=\"MODALITE\" )\n", "salaires.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On r\u00e9cup\u00e8re les variables utiles pour la r\u00e9gression."]}, {"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>AGE</th>\n", "      <th>M</th>\n", "      <th>F</th>\n", "      <th>montant</th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>0</th>\n", "      <td>49.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>1</th>\n", "      <td>27.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>2</th>\n", "      <td>22.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>3</th>\n", "      <td>26.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>4</th>\n", "      <td>29.0</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>750.0</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["    AGE  M  F  montant\n", "0  49.0  1  0    750.0\n", "1  27.0  1  0    750.0\n", "2  22.0  1  0    750.0\n", "3  26.0  1  0    750.0\n", "4  29.0  0  1    750.0"]}, "execution_count": 16, "metadata": {}, "output_type": "execute_result"}], "source": ["salaires[\"M\"] = salaires.apply( lambda r : 1 if r[\"SEXE\"] == \"1\" else 0, axis=1)\n", "salaires[\"F\"] = 1 - salaires[\"M\"]  # en supposant que le sexe est toujours renseign\u00e9\n", "data = salaires[[\"AGE\", \"M\", \"F\", \"montant\"]]\n", "data = data[data.M + data.F > 0]\n", "data.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Ce type d'\u00e9criture est plut\u00f4t lent car une fonction Python est ex\u00e9cut\u00e9e \u00e0 chaque it\u00e9ration. Il est pr\u00e9f\u00e9rable d\u00e8s que c'est possible d'utiliser les expressions avec des indices sans passer par la fonction [apply](http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.apply.html) qui cr\u00e9\u00e9 une copie de chaque ligne avant d'appliquer la fonction \u00e0 appliquer \u00e0 chacune d'entre elles."]}, {"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>TRNNETO</th>\n", "      <th>AGE</th>\n", "      <th>SEXE</th>\n", "      <th>VARIABLE</th>\n", "      <th>MODALITE</th>\n", "      <th>MODLIBELLE</th>\n", "      <th>montant</th>\n", "      <th>M</th>\n", "      <th>F</th>\n", "      <th>M2</th>\n", "      <th>F2</th>\n", "    </tr>\n", "  </thead>\n", "  <tbody>\n", "    <tr>\n", "      <th>0</th>\n", "      <td>02</td>\n", "      <td>49.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>1</th>\n", "      <td>02</td>\n", "      <td>27.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>2</th>\n", "      <td>02</td>\n", "      <td>22.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>3</th>\n", "      <td>02</td>\n", "      <td>26.0</td>\n", "      <td>1</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "    </tr>\n", "    <tr>\n", "      <th>4</th>\n", "      <td>02</td>\n", "      <td>29.0</td>\n", "      <td>2</td>\n", "      <td>TRNNETO</td>\n", "      <td>02</td>\n", "      <td>[500 ; 1 000[ euros</td>\n", "      <td>750.0</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "      <td>0</td>\n", "      <td>1</td>\n", "    </tr>\n", "  </tbody>\n", "</table>\n", "</div>"], "text/plain": ["  TRNNETO   AGE SEXE VARIABLE MODALITE           MODLIBELLE  montant  M  F  \\\n", "0      02  49.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0  1  0   \n", "1      02  27.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0  1  0   \n", "2      02  22.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0  1  0   \n", "3      02  26.0    1  TRNNETO       02  [500 ; 1 000[ euros    750.0  1  0   \n", "4      02  29.0    2  TRNNETO       02  [500 ; 1 000[ euros    750.0  0  1   \n", "\n", "   M2  F2  \n", "0   1   0  \n", "1   1   0  \n", "2   1   0  \n", "3   1   0  \n", "4   0   1  "]}, "execution_count": 17, "metadata": {}, "output_type": "execute_result"}], "source": ["salaires[\"M2\"] = 0\n", "salaires.loc[salaires[\"SEXE\"] == \"1\", \"M2\"] = 1\n", "salaires[\"F2\"] = 1 - salaires[\"M2\"]\n", "salaires.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Il ne reste plus qu'\u00e0 faire la r\u00e9gression."]}, {"cell_type": "markdown", "metadata": {}, "source": ["#### version scikit-learn\n", "\n", "Vous pouvez vous inspirer de cet [exemple](http://scikit-learn.org/stable/auto_examples/linear_model/plot_ols.html)."]}, {"cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": []}, {"cell_type": "markdown", "metadata": {}, "source": ["#### version statsmodels\n", "\n", "L'exemple avec ce module est [ici](http://statsmodels.sourceforge.net/devel/examples/notebooks/generated/ols.html)."]}, {"cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [], "source": []}], "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": "3.9.5"}}, "nbformat": 4, "nbformat_minor": 4}