{"id":9990,"date":"2020-09-22T12:33:11","date_gmt":"2020-09-22T12:33:11","guid":{"rendered":"https:\/\/datascientest.com\/?p=9990"},"modified":"2023-04-03T07:57:15","modified_gmt":"2023-04-03T07:57:15","slug":"challenge-data-science","status":"publish","type":"post","link":"https:\/\/datascientest.com\/challenge-data-science","title":{"rendered":"Enqu\u00eate Exclusive au coeur d&rsquo;un Data Challenge"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"9990\" class=\"elementor elementor-9990\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-6c452a0 e-flex e-con-boxed e-con\" data-id=\"6c452a0\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-348126c elementor-widget elementor-widget-heading\" data-id=\"348126c\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-heading-title{padding:0;margin:0;line-height:1}.elementor-widget-heading .elementor-heading-title[class*=elementor-size-]>a{color:inherit;font-size:inherit;line-height:inherit}.elementor-widget-heading .elementor-heading-title.elementor-size-small{font-size:15px}.elementor-widget-heading .elementor-heading-title.elementor-size-medium{font-size:19px}.elementor-widget-heading .elementor-heading-title.elementor-size-large{font-size:29px}.elementor-widget-heading .elementor-heading-title.elementor-size-xl{font-size:39px}.elementor-widget-heading .elementor-heading-title.elementor-size-xxl{font-size:59px}<\/style><h2 class=\"elementor-heading-title elementor-size-default\">Le 14 et 15 septembre dernier se tenait l\u2019\u00e9v\u00e9nement data de la rentr\u00e9e : le Salon Big Data Paris.\nDataScientest y \u00e9tait et avait \u00e9t\u00e9 s\u00e9lectionn\u00e9 pour l\u2019organisation d\u2019un data challenge. Au programme : exploiter les donn\u00e9es de plus de 400 000 stations m\u00e9t\u00e9orologiques pour mettre au point le meilleur mod\u00e8le de pr\u00e9diction de la concentration en particules fines. Retour sur une journ\u00e9e passionnante entre comp\u00e9tition et entraide intellectuelle.<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-8cb5b51 elementor-widget elementor-widget-text-editor\" data-id=\"8cb5b51\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-text-editor.elementor-drop-cap-view-stacked .elementor-drop-cap{background-color:#69727d;color:#fff}.elementor-widget-text-editor.elementor-drop-cap-view-framed .elementor-drop-cap{color:#69727d;border:3px solid;background-color:transparent}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap{margin-top:8px}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap-letter{width:1em;height:1em}.elementor-widget-text-editor .elementor-drop-cap{float:left;text-align:center;line-height:1;font-size:50px}.elementor-widget-text-editor .elementor-drop-cap-letter{display:inline-block}<\/style>\t\t\t\t<p>Chloe et Souhail, Data Scientist chez DataScientest ont pu encadrer les \u00e9quipes lors du challenge, sous la supervision de Charles Sutton &#8211; CTO de DataScientest. Ils nous racontent leur exp\u00e9rience :<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-f223b64 elementor-widget elementor-widget-image\" data-id=\"f223b64\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-image{text-align:center}.elementor-widget-image a{display:inline-block}.elementor-widget-image a img[src$=\".svg\"]{width:48px}.elementor-widget-image img{vertical-align:middle;display:inline-block}<\/style>\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"600\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-1024x768.jpg\" class=\"attachment-large size-large wp-image-10010\" alt=\"data challenge paris\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-1024x768.jpg 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-300x225.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-768x576.jpg 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-1536x1152.jpg 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-charlie-1-2048x1536.jpg 2048w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-7f00efa elementor-widget elementor-widget-text-editor\" data-id=\"7f00efa\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>\u00ab\u00a0Le Data Challenge organis\u00e9 par DataScientest au salon <a href=\"https:\/\/www.bigdataparis.com\/\"><span style=\"text-decoration: underline;\">BigDataParis<\/span><\/a>\u00a0\u00e9tait l\u2019occasion de simuler un <strong>environnement comp\u00e9titif<\/strong> <strong>sur le th\u00e8me de la Data Science<\/strong> pour de courtes sessions de 2h.\u00a0 L\u2019objectif \u00e9tait de r\u00e9soudre <strong>un probl\u00e8me de Machine Learning<\/strong> avec des donn\u00e9es directement disponibles sur la plateforme. Au programme du challenge : estimer le taux de NO\u00b2 en utilisant les donn\u00e9es des stations m\u00e9t\u00e9orologiques. Le challenge r\u00e9compensait le Data scientist pr\u00e9sentant un mod\u00e8le ayant l\u2019erreur la plus faible (RMSE).<\/p><p>Les comp\u00e9titeurs avaient tous un background de <strong>Data Science<\/strong>, mais des parcours et exp\u00e9riences diff\u00e9rentes. De ce fait, chacun a abord\u00e9 le challenge \u00e0 sa mani\u00e8re.<\/p><blockquote><p><em>M\u00eame si les performances \u00e9taient comptabilis\u00e9es individuellement certains ont pr\u00e9f\u00e9r\u00e9 <strong>travailler en groupe<\/strong>. Cette d\u00e9marche permettant aux <strong>d\u00e9butants<\/strong> de <strong>profiter de l\u2019exp\u00e9rience des plus experts<\/strong>. Le tout dans un <strong>climat d\u2019entraide<\/strong> <strong>et de stimulation<\/strong> li\u00e9e au challenge.<\/em>\u00a0<\/p><\/blockquote><p>Les 30 premi\u00e8res minutes sont globalement l\u2019occasion de se familiariser avec le probl\u00e8me et d\u2019<strong>appr\u00e9hender la plateforme de challenge<\/strong>. En milieu de session, <strong>les comp\u00e9titeurs explorent les donn\u00e9es fournies<\/strong> et se mettent \u00e0 <strong>construire leurs premiers mod\u00e8les<\/strong>.<\/p><p>\u00c0 mi-parcours, la pression commence \u00e0 monter d\u00e8s que des \u00e9quipes r\u00e9ussissent \u00e0 soumettre leur premiers r\u00e9sultats. En effet, un <strong>leaderboard<\/strong> est affich\u00e9 sur un \u00e9cran dans la salle de challenge et motive les \u00e9quipes \u00e0 soumettre leurs r\u00e9sultats pour <strong>am\u00e9liorer leur classement<\/strong>.\u00a0<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d0f18de elementor-widget elementor-widget-image\" data-id=\"d0f18de\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"600\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-1024x768.jpg\" class=\"attachment-large size-large wp-image-10012\" alt=\"leader board\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-1024x768.jpg 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-300x225.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-768x576.jpg 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-1536x1152.jpg 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/leader-board-1-2048x1536.jpg 2048w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-fc44b92 elementor-widget elementor-widget-text-editor\" data-id=\"fc44b92\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Ce syst\u00e8me de leaderboard et les enjeux du challenge (lots \u00e0 gagner) provoquent une <\/span><b>mont\u00e9e d\u2019adr\u00e9naline<\/b><span style=\"font-weight: 400;\"> vers le milieu\/fin de de session.\u00a0<\/span><\/p><blockquote><p><em><span style=\"font-weight: 400;\">Pour chaque session on a pu observer en moyenne <\/span><b>10 mod\u00e8les soumis <\/b><span style=\"font-weight: 400;\">par participants dans la <\/span><b>derni\u00e8re heure <\/b><\/em><span style=\"font-weight: 400;\"><em>du challenge.<\/em>\u00a0<\/span><\/p><\/blockquote><p><span style=\"font-weight: 400;\">Certains n\u2019h\u00e9sitent plus \u00e0 <\/span><b>poser plusieurs questions aux coach pour les aider \u00e0 avancer<\/b><span style=\"font-weight: 400;\">, d\u2019autres s\u2019amusent \u00e0 <\/span><b>surveiller leur position sur le leaderboard<\/b><span style=\"font-weight: 400;\"> et \u00e0 soumettre plusieurs fois des r\u00e9sultats. On sent que les \u00e9quipes prennent tr\u00e8s au s\u00e9rieux ce challenge ce qui am\u00e9liore leur exp\u00e9rience.<\/span><\/p><p><span style=\"font-weight: 400;\">A la fin de session, la pression retombe et les r\u00e9sultats finaux sont affich\u00e9s. Quelque soit le classement final, tous s\u2019accordent sur le fait que c\u2019\u00e9tait une tr\u00e8s bonne exp\u00e9rience et n\u2019h\u00e9sitent pas \u00e0 nous parler de la <\/span><b>proc\u00e9dure<\/b><span style=\"font-weight: 400;\"> et <\/span><b>m\u00e9thodologie<\/b><span style=\"font-weight: 400;\"> qu\u2019ils ont suivi.\u00a0<\/span><\/p><blockquote><p><em><span style=\"font-weight: 400;\">La courte dur\u00e9e du challenge a permis aux comp\u00e9titeurs de<\/span><b> travailler avec intensit\u00e9 et concentration<\/b><span style=\"font-weight: 400;\"> et nul doute que cela leur servira beaucoup dans leur parcours en Data Science.\u00a0\u00bb<\/span><\/em><\/p><\/blockquote><p><span style=\"font-weight: 400;\">Nous avons \u00e9galement demand\u00e9 \u00e0 certains des comp\u00e9titeurs dont la performance leur a permis d\u2019atteindre le podium de cette comp\u00e9tition de nous pr\u00e9senter leurs approches respectives :<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-c176a92 e-flex e-con-boxed e-con\" data-id=\"c176a92\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-bacf6d6 elementor-widget elementor-widget-heading\" data-id=\"bacf6d6\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">1 ) Amed Coulibaly-Data Scientist chez Metro France<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-074a34e elementor-widget elementor-widget-text-editor\" data-id=\"074a34e\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">\u00ab\u00a0Ma solution comporte trois parties importantes :<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-90678de elementor-widget elementor-widget-heading\" data-id=\"90678de\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">a) Feature engineering :<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d40a58a elementor-widget elementor-widget-text-editor\" data-id=\"d40a58a\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Les variables explicatives (donn\u00e9es climatiques, position des stations &#8230;) \u00e9taient d&rsquo;ordre temporel, il \u00e9tait donc important de cr\u00e9er des features comme ci-dessous.<\/p><ul><li>Les <strong>Lags<\/strong> (des fen\u00eatres de valeurs d\u00e9cal\u00e9es de chaque features).<\/li><li><strong>Rolling Means<\/strong> (des fen\u00eatres de moyennes mobiles sur les lags) ce qui permet de capter la tendance des features.<\/li><\/ul>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-b78a91b elementor-widget elementor-widget-heading\" data-id=\"b78a91b\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">b) Strat\u00e9gie de CrossValidation :<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-a398836 elementor-widget elementor-widget-text-editor\" data-id=\"a398836\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Avec les s\u00e9ries temporelles il est souvent tr\u00e8s important d&rsquo;adapter sa strat\u00e9gie de validation en fonction du jeu de test de telle sorte que le mod\u00e8le soit assez g\u00e9n\u00e9ralisable. Pour ce faire, j&rsquo;ai donc utilis\u00e9 un groupkfold split en fonction des stations et du trimestre.<\/p><ul><li>Modeling :<\/li><\/ul><p style=\"padding-left: 40px;\">J&rsquo;ai utilis\u00e9 un mod\u00e8le de <strong>boosting<\/strong> (lightgbm).<\/p><ul><li>Blending :<\/li><\/ul><p style=\"padding-left: 40px;\">Avec le ligthgbm en changeant ci-dessus la strat\u00e9gie de validation ou les param\u00e8tres du lgbm on obtient diff\u00e9rent pr\u00e9dicteurs. Mon id\u00e9e \u00e9tait de combiner les d\u00e9cisions du plus grand nombre de pr\u00e9dicteurs possible.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d6928fe elementor-widget elementor-widget-text-editor\" data-id=\"d6928fe\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Sur ce, en 2h j&rsquo;ai pu faire 11 exp\u00e9rimentations de 5 folds chacune qui se r\u00e9sument en 55 pr\u00e9dicteurs. Mon mod\u00e8le final est une <strong>combinaison de la d\u00e9cision de ces pr\u00e9dicteurs<\/strong>.<\/p><p>Si j&rsquo;avais eu plus de temps, j&rsquo;aurais certainement impl\u00e9ment\u00e9 une iterative pr\u00e9diction, qui consiste \u00e0 <strong>pr\u00e9dire it\u00e9rativement la target<\/strong> et non toutes les valeurs d&rsquo;un seul coup. Cela permet au mod\u00e8le d&rsquo;avoir de l&rsquo;information sur les valeurs pr\u00e9c\u00e9demment pr\u00e9dites et d&rsquo;am\u00e9liorer ses pr\u00e9dictions suivantes.\u00a0\u00bb<\/p><p><strong>RMSE : 8,8092<\/strong><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2a1a8c9 elementor-widget elementor-widget-image\" data-id=\"2a1a8c9\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"600\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-1024x768.jpg\" class=\"attachment-large size-large wp-image-10024\" alt=\"data challenge\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-1024x768.jpg 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-300x225.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-768x576.jpg 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-1536x1152.jpg 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-equipe-coding-1-2048x1536.jpg 2048w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-6847bec e-flex e-con-boxed e-con\" data-id=\"6847bec\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-331b023 elementor-widget elementor-widget-heading\" data-id=\"331b023\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">2) Thomas Denimal - Data Scientist chez Lyreco Group<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-29bb7c7 elementor-widget elementor-widget-heading\" data-id=\"29bb7c7\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">a) Mod\u00e8le utilis\u00e9<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-983b3fd elementor-widget elementor-widget-text-editor\" data-id=\"983b3fd\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>\u00ab\u00a0D\u00e8s le d\u00e9part, je voulais utiliser la biblioth\u00e8que lightgbm: rapide, elle g\u00e8re de nombreux cas d&rsquo;utilisation diff\u00e9rents et fournit d&rsquo;excellents r\u00e9sultats. J&rsquo;ai apport\u00e9 des modifications aux param\u00e8tres du mod\u00e8le (en augmentant num_leaves, num_iterations) avec des am\u00e9liorations progressives.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-7595dc6 elementor-widget elementor-widget-heading\" data-id=\"7595dc6\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-default\">b) Fonctionnalit\u00e9s utilis\u00e9es<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-01641bb elementor-widget elementor-widget-text-editor\" data-id=\"01641bb\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">J&rsquo;ai commenc\u00e9 le d\u00e9fi en utilisant uniquement la mesure de NO\u00b2 de la station la plus proche et j&rsquo;ai introduit les features de la 2e station la plus proche \u00e0 la 5e, une \u00e0 la fois. Une<\/span><b> validation crois\u00e9e <\/b><span style=\"font-weight: 400;\">\u00e0 10 folds a \u00e9t\u00e9 utilis\u00e9e pour valider le mod\u00e8le. \u00c0 la fin, j&rsquo;ai \u00e9galement ajout\u00e9 la <\/span><b>valeur moyenne des mesures de NO\u00b2 <\/b><span style=\"font-weight: 400;\">qui a \u00e9t\u00e9 utilis\u00e9e dans la m\u00e9thode na\u00efve avec des am\u00e9liorations visibles de la validation crois\u00e9e\u00a0 (RMSE moyen et std-standard deviation- du rmse moyen) et de mon classement \ud83d\ude42\u00a0<\/span><\/p><blockquote><p><em><span style=\"font-weight: 400;\">Le trac\u00e9 du RMSE moyen et du std du RMSE moyen fournit des informations utiles pour d\u00e9tecter le sous-ajustement \/ le surajustement et les changements de param\u00e8tres.<\/span><\/em><\/p><\/blockquote><p><span style=\"font-weight: 400;\">J&rsquo;ai essay\u00e9 d&rsquo;ajouter d&rsquo;autres statistiques, comme les mesures min, max et std de NO\u00b2, mais je n&rsquo;ai pas remarqu\u00e9 d&rsquo;am\u00e9liorations.\u00a0\u00bb<\/span><\/p><p><b>RMSE : 8,8560<\/b><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-6cf1d3a e-flex e-con-boxed e-con\" data-id=\"6cf1d3a\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-cbd1ace elementor-widget elementor-widget-heading\" data-id=\"cbd1ace\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">3)\u00a0Thierry\u00a0GAMEIRO-MARTINS - Data Scientist au Minist\u00e8re de l\u2019Economie et des Finances publiques<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-571b55a elementor-widget elementor-widget-text-editor\" data-id=\"571b55a\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">\u00ab\u00a0Ma m\u00e9thode consistait \u00e0 essayer quelques mod\u00e8les sur l&rsquo;ensemble des donn\u00e9es puis de seulement s\u00e9lectionner\u00a0 certaines variables. J&rsquo;ai notamment utilis\u00e9 un random forest, un xgboost et lightgbm.<\/span><\/p><p><span style=\"font-weight: 400;\">Pour\u00a0<\/span><a href=\"https:\/\/datascientest.com\/xgboost-grand-gagnant-des-competitions-machine-learning-algorithme\" target=\"_blank\" rel=\"noopener\">en savoir plus sur l\u2019algorithme xgboost<\/a><span style=\"font-weight: 400;\">,\u00a0<\/span><a href=\"https:\/\/datascientest.com\/xgboost-grand-gagnant-des-competitions-machine-learning-algorithme\" target=\"_blank\" rel=\"noopener\">d\u00e9couvrez cet article<\/a><\/p><p><span style=\"font-weight: 400;\">J&rsquo;ai ensuite tent\u00e9 un peu de feature engineering par de la discr\u00e9tisation de variables.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Enfin, j&rsquo;ai voulu me tourner vers l&rsquo;ajout de donn\u00e9es &#8211; temporelle, et de localisation disponible gr\u00e2ce aux identifiants- mais sans pouvoir terminer.\u00a0\u00bb<\/span><\/p><p><b>RMSE: 8,9429<\/b><\/p><p>Finalement, ce data challenge a permis de\u00a0<strong>r\u00e9unir des talents d\u2019horizons diff\u00e9rents<\/strong>\u00a0autour d\u2019une m\u00eame\u00a0<strong>probl\u00e9matique\u00a0<\/strong>et de la r\u00e9soudre \u00e0 l\u2019aide de la\u00a0<strong>data science<\/strong>.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-bd56465 elementor-widget elementor-widget-image\" data-id=\"bd56465\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"768\" height=\"456\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-768x456.jpg\" class=\"attachment-medium_large size-medium_large wp-image-10140\" alt=\"data challenge team\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-768x456.jpg 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-300x178.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-1024x609.jpg 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-1536x913.jpg 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/data-challenge-team-2048x1217.jpg 2048w\" sizes=\"(max-width: 768px) 100vw, 768px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-50310aa e-flex e-con-boxed e-con\" data-id=\"50310aa\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-47ba406 elementor-widget__width-initial elementor-widget elementor-widget-video\" data-id=\"47ba406\" data-element_type=\"widget\" data-settings=\"{&quot;youtube_url&quot;:&quot;https:\\\/\\\/youtu.be\\\/NpCtTJtVXxI&quot;,&quot;video_type&quot;:&quot;youtube&quot;,&quot;controls&quot;:&quot;yes&quot;}\" data-widget_type=\"video.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-video .elementor-widget-container{overflow:hidden;transform:translateZ(0)}.elementor-widget-video .elementor-wrapper{aspect-ratio:var(--video-aspect-ratio)}.elementor-widget-video .elementor-wrapper iframe,.elementor-widget-video .elementor-wrapper video{height:100%;width:100%;display:flex;border:none;background-color:#000}@supports not (aspect-ratio:1\/1){.elementor-widget-video .elementor-wrapper{position:relative;overflow:hidden;height:0;padding-bottom:calc(100% \/ var(--video-aspect-ratio))}.elementor-widget-video .elementor-wrapper iframe,.elementor-widget-video .elementor-wrapper video{position:absolute;top:0;right:0;bottom:0;left:0}}.elementor-widget-video .elementor-open-inline .elementor-custom-embed-image-overlay{position:absolute;top:0;right:0;bottom:0;left:0;background-size:cover;background-position:50%}.elementor-widget-video .elementor-custom-embed-image-overlay{cursor:pointer;text-align:center}.elementor-widget-video .elementor-custom-embed-image-overlay:hover .elementor-custom-embed-play i{opacity:1}.elementor-widget-video .elementor-custom-embed-image-overlay img{display:block;width:100%;aspect-ratio:var(--video-aspect-ratio);-o-object-fit:cover;object-fit:cover;-o-object-position:center center;object-position:center center}@supports not (aspect-ratio:1\/1){.elementor-widget-video .elementor-custom-embed-image-overlay{position:relative;overflow:hidden;height:0;padding-bottom:calc(100% \/ var(--video-aspect-ratio))}.elementor-widget-video .elementor-custom-embed-image-overlay img{position:absolute;top:0;right:0;bottom:0;left:0}}.elementor-widget-video .e-hosted-video .elementor-video{-o-object-fit:cover;object-fit:cover}.e-con-inner>.elementor-widget-video,.e-con>.elementor-widget-video{width:var(--container-widget-width);--flex-grow:var(--container-widget-flex-grow)}<\/style>\t\t<div class=\"elementor-wrapper elementor-open-inline\">\n\t\t\t<div class=\"elementor-video\"><\/div>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-5fecdd0 e-flex e-con-boxed e-con\" data-id=\"5fecdd0\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-d91efa0 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"d91efa0\" data-element_type=\"widget\" data-widget_type=\"button.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-button-wrapper\">\n\t\t\t<a class=\"elementor-button elementor-button-link elementor-size-sm elementor-animation-grow\" href=\"https:\/\/datascientest.com\/data-challenge\">\n\t\t\t\t\t\t<span class=\"elementor-button-content-wrapper\">\n\t\t\t\t\t\t<span class=\"elementor-button-text\">En savoir plus sur le Data Challenge<\/span>\n\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Le 14 et 15 septembre dernier se tenait l\u2019\u00e9v\u00e9nement data de la rentr\u00e9e : le Salon Big Data Paris. DataScientest y \u00e9tait et avait \u00e9t\u00e9 s\u00e9lectionn\u00e9 pour l\u2019organisation d\u2019un data challenge. Au programme : exploiter les donn\u00e9es de plus de 400 000 stations m\u00e9t\u00e9orologiques pour mettre au point le meilleur mod\u00e8le de pr\u00e9diction de la [&hellip;]<\/p>\n","protected":false},"author":45,"featured_media":10102,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_theme","format":"standard","meta":{"content-type":"","inline_featured_image":false},"categories":[2431],"tags":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v21.7 (Yoast SEO v21.7) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Data Challenge : Enqu\u00eate au coeur de notre dispositif !<\/title>\n<meta name=\"description\" content=\"Et si DataScientest vous faisait vivre une comp\u00e9tition de data science comme si vous y \u00e9tiez ? Retour sur un data challenge d&#039;exception !\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/datascientest.com\/challenge-data-science\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Enqu\u00eate Exclusive au coeur d&#039;un Data Challenge\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/challenge-data-science\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2020-09-22T12:33:11+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-04-03T07:57:15+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i0.wp.com\/datascientest.com\/wp-content\/uploads\/2020\/09\/BIGData_Article-01.png?fit=2560%2C1010&ssl=1\" \/>\n\t<meta property=\"og:image:width\" content=\"2560\" \/>\n\t<meta property=\"og:image:height\" content=\"1010\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Daniel\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Daniel\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"7 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/challenge-data-science\",\"url\":\"https:\/\/datascientest.com\/challenge-data-science\",\"name\":\"Data Challenge : Enqu\u00eate au coeur de notre dispositif !\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2020-09-22T12:33:11+00:00\",\"dateModified\":\"2023-04-03T07:57:15+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/5a86e30c5090234fb43dd8baa83c44b0\"},\"description\":\"Et si DataScientest vous faisait vivre une comp\u00e9tition de data science comme si vous y \u00e9tiez ? Retour sur un data challenge d'exception !\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/challenge-data-science\"]}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/datascientest.com\/#website\",\"url\":\"https:\/\/datascientest.com\/\",\"name\":\"Formation Data Science | DataScientest.com\",\"description\":\"Leader de la formation en Data Science\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/datascientest.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/5a86e30c5090234fb43dd8baa83c44b0\",\"name\":\"Daniel\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/10\/dst.svg\",\"contentUrl\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/10\/dst.svg\",\"caption\":\"Daniel\"},\"sameAs\":[\"https:\/\/www.linkedin.com\/school\/datascientest\/\",\"user\"],\"jobTitle\":\"ing\u00e9nieur logiciel\",\"worksFor\":\"DataScientest\",\"url\":\"https:\/\/datascientest.com\/author\/team-redac\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Data Challenge : Enqu\u00eate au coeur de notre dispositif !","description":"Et si DataScientest vous faisait vivre une comp\u00e9tition de data science comme si vous y \u00e9tiez ? Retour sur un data challenge d'exception !","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/datascientest.com\/challenge-data-science","og_locale":"fr_FR","og_type":"article","og_title":"Enqu\u00eate Exclusive au coeur d'un Data Challenge","og_url":"https:\/\/datascientest.com\/challenge-data-science","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2020-09-22T12:33:11+00:00","article_modified_time":"2023-04-03T07:57:15+00:00","og_image":[{"width":2560,"height":1010,"url":"https:\/\/i0.wp.com\/datascientest.com\/wp-content\/uploads\/2020\/09\/BIGData_Article-01.png?fit=2560%2C1010&ssl=1","type":"image\/png"}],"author":"Daniel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Daniel","Dur\u00e9e de lecture estim\u00e9e":"7 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/challenge-data-science","url":"https:\/\/datascientest.com\/challenge-data-science","name":"Data Challenge : Enqu\u00eate au coeur de notre dispositif !","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2020-09-22T12:33:11+00:00","dateModified":"2023-04-03T07:57:15+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/5a86e30c5090234fb43dd8baa83c44b0"},"description":"Et si DataScientest vous faisait vivre une comp\u00e9tition de data science comme si vous y \u00e9tiez ? Retour sur un data challenge d'exception !","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/challenge-data-science"]}]},{"@type":"WebSite","@id":"https:\/\/datascientest.com\/#website","url":"https:\/\/datascientest.com\/","name":"Formation Data Science | DataScientest.com","description":"Leader de la formation en Data Science","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/datascientest.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"fr-FR"},{"@type":"Person","@id":"https:\/\/datascientest.com\/#\/schema\/person\/5a86e30c5090234fb43dd8baa83c44b0","name":"Daniel","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/datascientest.com\/#\/schema\/person\/image\/","url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/10\/dst.svg","contentUrl":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/10\/dst.svg","caption":"Daniel"},"sameAs":["https:\/\/www.linkedin.com\/school\/datascientest\/","user"],"jobTitle":"ing\u00e9nieur logiciel","worksFor":"DataScientest","url":"https:\/\/datascientest.com\/author\/team-redac"}]}},"_links":{"self":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/9990"}],"collection":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/users\/45"}],"replies":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/comments?post=9990"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/9990\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/10102"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=9990"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=9990"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=9990"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}