FlorianSC/agritech-interface

1{2 "cells": [3  {4   "cell_type": "markdown",5   "id": "50a716c8",6   "metadata": {},7   "source": [8    "# Système de recommandation agricole - Feature engineering\n",9    "- L'objectif ici va d'être de réaliser l'enrichissement de notre nouveau df avec les varibales clés de crop_yield\n",10    "- Ensuite on va passer au feature engineering à proprement parlé"11   ]12  },13  {14   "cell_type": "code",15   "execution_count": 2,16   "id": "751f253d",17   "metadata": {},18   "outputs": [],19   "source": [20    "# Imports de base\n",21    "import pandas as pd\n",22    "import numpy as np\n",23    "import seaborn as sns\n",24    "import matplotlib.pyplot as plt\n",25    "import scipy.stats as stats"26   ]27  },28  {29   "cell_type": "markdown",30   "id": "c2304a35",31   "metadata": {},32   "source": [33    "## On charge nos données"34   ]35  },36  {37   "cell_type": "code",38   "execution_count": 2,39   "id": "8573196b",40   "metadata": {},41   "outputs": [],42   "source": [43    "# Fichier yield consolidé\n",44    "data_yield_conso = pd.read_csv(\"../data/processed/yield_df_final.csv\")\n",45    "crop_yield = pd.read_csv(\"../data/raw/crop_yield.csv\")"46   ]47  },48  {49   "cell_type": "markdown",50   "id": "856dbfd3",51   "metadata": {},52   "source": [53    "### Vérification des cultures par fichier"54   ]55  },56  {57   "cell_type": "code",58   "execution_count": 3,59   "id": "3836060d",60   "metadata": {},61   "outputs": [62    {63     "name": "stdout",64     "output_type": "stream",65     "text": [66      "Type de culture dans yield :<StringArray>\n",67      "[               'Maize',             'Potatoes',          'Rice, paddy',\n",68      "                'Wheat',              'Sorghum',             'Soybeans',\n",69      "              'Cassava',                 'Yams',       'Sweet potatoes',\n",70      " 'Plantains and others']\n",71      "Length: 10, dtype: str\n",72      "\n",73      "Type de culture dans crop_yield :<StringArray>\n",74      "['Cotton', 'Rice', 'Barley', 'Soybean', 'Wheat', 'Maize']\n",75      "Length: 6, dtype: str\n",76      "\n",77      "====================\n",78      "Fichier yield consolidé : item\n",79      "Maize                   4426\n",80      "Potatoes                4163\n",81      "Wheat                   3289\n",82      "Rice, paddy             3171\n",83      "Sweet potatoes          3111\n",84      "Sorghum                 2826\n",85      "Cassava                 2771\n",86      "Soybeans                2428\n",87      "Yams                    1615\n",88      "Plantains and others    1351\n",89      "Name: count, dtype: int64\n",90      "\n",91      "Fichier crop_yield : Crop\n",92      "Maize      166824\n",93      "Rice       166792\n",94      "Barley     166777\n",95      "Wheat      166673\n",96      "Cotton     166585\n",97      "Soybean    166349\n",98      "Name: count, dtype: int64\n"99     ]100    }101   ],102   "source": [103    "# Type de culture unique par fichier\n",104    "print(f\"Type de culture dans yield :{data_yield_conso['item'].unique()}\\n\")\n",105    "print(f\"Type de culture dans crop_yield :{crop_yield['Crop'].unique()}\\n\")\n",106    "print(\"=\"*20)\n",107    "# Valeur par type de culture par fichier\n",108    "print(f\"Fichier yield consolidé : {data_yield_conso['item'].value_counts()}\\n\")\n",109    "print(f\"Fichier crop_yield : {crop_yield['Crop'].value_counts()}\")"110   ]111  },112  {113   "cell_type": "markdown",114   "id": "021fd1f6",115   "metadata": {},116   "source": [117    "On note une différence entre les fichiers sur les types de culture :\n",118    "\n",119    "- Nous devons effectuer un mapping pour faire correspondre les bonnes cultures\n",120    "- **On note également qu'il y a 10 cultures dans le fchier consolidé et seulement 6 dans le dataset synthétique. Il faudrait bien avoir ça en tête lors de l'imputation.**\n",121    "- Les éléments suivants n'auront pas d'équivalents directs :\n",122    "    - Potatoes\n",123    "    - Sweet potatoes\n",124    "    - Sorghum\n",125    "    - Cassava\n",126    "    - Yams\n",127    "    - Plantains and others\n",128    "- L'idée va être de réunir un maximum d'informations depuis la relation culture des deux datasets et pour les relations directes manquantes, nous allons le faire par pays dans le fichier consolidé puis par groupe climatique\n",129    "- On voit également que la dataset \"synthétique\" propose un équilibre quasi parfait entre type de culture"130   ]131  },132  {133   "cell_type": "markdown",134   "id": "063d18f5",135   "metadata": {},136   "source": [137    "### Analyse des variables similaires"138   ]139  },140  {141   "cell_type": "code",142   "execution_count": 4,143   "id": "6e119d9a",144   "metadata": {},145   "outputs": [146    {147     "name": "stdout",148     "output_type": "stream",149     "text": [150      "Fichier yield consolidé : \n",151      "           avg_temp   rainfall_mm\n",152      "count  29151.000000  29151.000000\n",153      "mean      20.286183   1227.017289\n",154      "std        5.911377    746.155030\n",155      "min       -3.370000     51.000000\n",156      "25%       17.330000    637.000000\n",157      "50%       21.340000   1146.000000\n",158      "75%       24.930000   1705.000000\n",159      "max       30.420000   3240.000000\n",160      "\n",161      "Fichier crop_yield : \n",162      "       Temperature_Celsius     Rainfall_mm\n",163      "count       1000000.000000  1000000.000000\n",164      "mean             27.504965      549.981901\n",165      "std               7.220608      259.851320\n",166      "min              15.000034      100.000896\n",167      "25%              21.254502      324.891090\n",168      "50%              27.507365      550.124061\n",169      "75%              33.753267      774.738520\n",170      "max              39.999997      999.998098\n",171      "\n"172     ]173    }174   ],175   "source": [176    "print(f\"Fichier yield consolidé : \\n{data_yield_conso[[\"avg_temp\",\"rainfall_mm\"]].describe()}\\n\")\n",177    "print(f\"Fichier crop_yield : \\n{crop_yield[[\"Temperature_Celsius\",\"Rainfall_mm\"]].describe()}\\n\")"178   ]179  },180  {181   "cell_type": "markdown",182   "id": "9ca907de",183   "metadata": {},184   "source": [185    "Les statistiques ne donnent pas les mêmes amplitudes par variables :\n",186    "- Du côté du fichier consolidé sur la temp on a pas la distinction des pays avec climat froid, le minimum est à 15 degrés alors qu'il est à -3 pour le fichier consolidé, on note un écart pour le maximum également.\n",187    "- Ce n'est pas parfait pour ajouter des données mais l'idée est de rajouter des proxys, donc des mesures approximatives afin d'aider un modèle à trouver des patterns."188   ]189  },190  {191   "cell_type": "markdown",192   "id": "03587d06",193   "metadata": {},194   "source": [195    "### Normalisation des noms dans notre fichier consolidé"196   ]197  },198  {199   "cell_type": "code",200   "execution_count": 5,201   "id": "73a7aa28",202   "metadata": {},203   "outputs": [204    {205     "data": {206      "text/plain": [207       "<StringArray>\n",208       "[               'Maize',             'Potatoes',                 'Rice',\n",209       "                'Wheat',              'Sorghum',              'Soybean',\n",210       "              'Cassava',                 'Yams',       'Sweet potatoes',\n",211       " 'Plantains and others']\n",212       "Length: 10, dtype: str"213      ]214     },215     "execution_count": 5,216     "metadata": {},217     "output_type": "execute_result"218    }219   ],220   "source": [221    "# Utilisation de replace pour changer les noms\n",222    "data_yield_conso['item'] = data_yield_conso['item'].replace({\n",223    "    'Rice, paddy': 'Rice',\n",224    "    'Soybeans': 'Soybean',\n",225    "    'Maize': 'Maize',\n",226    "    'Wheat': 'Wheat',\n",227    "    'Barley': 'Barley'\n",228    "})\n",229    "data_yield_conso['item'].unique()"230   ]231  },232  {233   "cell_type": "markdown",234   "id": "d96e40d4",235   "metadata": {},236   "source": [237    "## Enrichissement du fichier consolidé\n",238    "\n",239    "Il n'y a de relation directe entre notre premier dataset (celui pour l'ACP) et les autres datasets.\n",240    "- Le premier fichier est par grande région et culture\n",241    "- Le deuxième fichier est par pays, par année puis par culture\n",242    "- Le rapprochement des données serait très bancal.\n",243    "\n",244    "**Mais nous avons identifié des variables proxy qui pourraient enrichir notre dataset :**\n",245    "- Irrigation_Used\n",246    "- Fertilizer_Used\n",247    "- Weather_Condition\n",248    "- Soil_Type\n",249    "\n",250    "Le fichier crop_yield semble être un fichier synthétique :\n",251    "- Pas de corrélation entre les variables explicatives\n",252    "- Répartition à quasi 50/50 des valeurs pour chaque variables \n",253    "\n",254    "Plusieurs méthodes d'enrichissement ont été effectué :\n",255    "- Modèle de régression logistique afin de prédire les différentes variables sur le nouveau fichier\n",256    "    - Variables explicative : précipitation et température\n",257    "    - Après un test, le modèle n'arrive pas à généraliser, les variables explicatives sont indépendantes entre elles Les données sont \"plates\" (sans corrélations fortes), le modèle de ML essayait de trouver une règle là où il n'y en a pas\n",258    "    - Par exemple, la balanced_accuracy était à 0,50 pour le fertilisant et l'irrigation\n",259    "\n",260    "- Méthode de stratification par catégorie afin de faire le rapprochement sur le nouveau fichier en utilisation la valeur la plus fréquente (mode())\n",261    "    - La mise en place de la stratification était simple (en quartiles)\n",262    "    - Le problème est l'utilisation de la valeur fréquente. Quand il y a une valeur qui se démarque, cela fonctionne bien. Cependant sur une égalité, le mode choisit aléatoirement\n",263    "- Méthode de sampling conditionnel afin d'attribuer la distribution de crop_yield par culture/clé climatique puis choix statistique de la valeur avec une fonction random\n",264    "    - Méthode retenue"265   ]266  },267  {268   "cell_type": "markdown",269   "id": "0d3d5a9a",270   "metadata": {},271   "source": [272    "### Méthode du sampling conditionnel\n",273    "\n",274    "- Standardisation des noms des types de culture entre les fichiers\n",275    "- Autre problématique : nous avons uniquement 4 cultures en communs entre les 2 fichiers"276   ]277  },278  {279   "cell_type": "code",280   "execution_count": 6,281   "id": "fa5c63a9",282   "metadata": {},283   "outputs": [],284   "source": [285    "# Avoir de pouvoir reproduire nos éléments\n",286    "np.random.seed(42)\n",287    "# Chargement de nos fichiers\n",288    "yield_df_final = data_yield_conso.copy()\n",289    "yield_df = yield_df_final.copy()\n",290    "crop_df = crop_yield.copy()\n",291    "# Colonnes que l'on veut retrouver\n",292    "cols = [\"Fertilizer_Used\", \"Irrigation_Used\", \"Weather_Condition\", \"Soil_Type\"]\n",293    "# Créer les colonnes manquantes en object, pas en float\n",294    "for col in cols:\n",295    "    if col not in yield_df.columns:\n",296    "        yield_df[col] = pd.Series([None] * len(yield_df), dtype=\"object\")\n",297    "    else:\n",298    "        yield_df[col] = yield_df[col].astype(\"object\")"299   ]300  },301  {302   "cell_type": "markdown",303   "id": "2b56342a",304   "metadata": {},305   "source": [306    "### Création de la clé climatique"307   ]308  },309  {310   "cell_type": "code",311   "execution_count": 7,312   "id": "6a5ae0e9",313   "metadata": {},314   "outputs": [],315   "source": [316    "# Construire climate_key avec les bornes apprises sur crop_yield\n",317    "temp_edges = crop_df[\"Temperature_Celsius\"].quantile([0.25, 0.5, 0.75]).tolist()\n",318    "rain_edges = crop_df[\"Rainfall_mm\"].quantile([0.25, 0.5, 0.75]).tolist()\n",319    "\n",320    "temp_bins = [-np.inf] + temp_edges + [np.inf]\n",321    "rain_bins = [-np.inf] + rain_edges + [np.inf]\n",322    "\n",323    "labels_t = [\"T1\", \"T2\", \"T3\", \"T4\"]\n",324    "labels_r = [\"R1\", \"R2\", \"R3\", \"R4\"]\n",325    "\n",326    "crop_df[\"temp_q\"] = pd.cut(\n",327    "    crop_df[\"Temperature_Celsius\"],\n",328    "    bins=temp_bins,\n",329    "    labels=labels_t,\n",330    "    include_lowest=True\n",331    ")\n",332    "crop_df[\"rain_q\"] = pd.cut(\n",333    "    crop_df[\"Rainfall_mm\"],\n",334    "    bins=rain_bins,\n",335    "    labels=labels_r,\n",336    "    include_lowest=True\n",337    ")\n",338    "crop_df[\"climate_key\"] = crop_df[\"temp_q\"].astype(str) + \"_\" + crop_df[\"rain_q\"].astype(str)\n",339    "\n",340    "yield_df[\"temp_q\"] = pd.cut(\n",341    "    yield_df[\"avg_temp\"],\n",342    "    bins=temp_bins,\n",343    "    labels=labels_t,\n",344    "    include_lowest=True\n",345    ")\n",346    "yield_df[\"rain_q\"] = pd.cut(\n",347    "    yield_df[\"rainfall_mm\"],\n",348    "    bins=rain_bins,\n",349    "    labels=labels_r,\n",350    "    include_lowest=True\n",351    ")\n",352    "yield_df[\"climate_key\"] = yield_df[\"temp_q\"].astype(str) + \"_\" + yield_df[\"rain_q\"].astype(str)\n"353   ]354  },355  {356   "cell_type": "markdown",357   "id": "5bf7c4c2",358   "metadata": {},359   "source": [360    "### Récupération de la distribution des éléments depuis crop_yield"361   ]362  },363  {364   "cell_type": "code",365   "execution_count": 8,366   "id": "5779a19c",367   "metadata": {},368   "outputs": [],369   "source": [370    "# Distributions depuis crop_yield\n",371    "dist_ref = {}\n",372    "dist_crop = {}\n",373    "dist_global = {}\n",374    "\n",375    "for col in cols:\n",376    "    dist_ref[col] = crop_df.groupby([\"climate_key\", \"Crop\"])[col].value_counts(normalize=True)\n",377    "    dist_crop[col] = crop_df.groupby(\"Crop\")[col].value_counts(normalize=True)\n",378    "    dist_global[col] = crop_df[col].value_counts(normalize=True)\n",379    "\n",380    "# Fonction de sampling\n",381    "def sample_value(row, col):\n",382    "    ck = row[\"climate_key\"]\n",383    "    crop = row[\"item\"]\n",384    "\n",385    "    # climate_key + crop\n",386    "    try:\n",387    "        dist = dist_ref[col].loc[(ck, crop)]\n",388    "        return np.random.choice(dist.index.to_list(), p=dist.values)\n",389    "    except Exception:\n",390    "        pass\n",391    "\n",392    "    # crop seul\n",393    "    try:\n",394    "        dist = dist_crop[col].loc[crop]\n",395    "        return np.random.choice(dist.index.to_list(), p=dist.values)\n",396    "    except Exception:\n",397    "        pass\n",398    "\n",399    "    # global\n",400    "    dist = dist_global[col]\n",401    "    return np.random.choice(dist.index.to_list(), p=dist.values)"402   ]403  },404  {405   "cell_type": "markdown",406   "id": "8170fde3",407   "metadata": {},408   "source": [409    "### Imputation des NaN ligne par ligne"410   ]411  },412  {413   "cell_type": "code",414   "execution_count": 9,415   "id": "68f6586c",416   "metadata": {},417   "outputs": [418    {419     "name": "stdout",420     "output_type": "stream",421     "text": [422      "NaN restants :\n",423      "Fertilizer_Used      0\n",424      "Irrigation_Used      0\n",425      "Weather_Condition    0\n",426      "Soil_Type            0\n",427      "dtype: int64\n"428     ]429    }430   ],431   "source": [432    "# Remplir les NaN ligne par ligne\n",433    "for col in cols:\n",434    "    mask = yield_df[col].isna()\n",435    "\n",436    "    if mask.sum() > 0:\n",437    "        sampled_values = yield_df.loc[mask].apply(\n",438    "            lambda row: sample_value(row, col),\n",439    "            axis=1\n",440    "        )\n",441    "\n",442    "        yield_df.loc[mask, col] = sampled_values.astype(\"object\")\n",443    "\n",444    "# Recast final propre\n",445    "yield_df[\"Fertilizer_Used\"] = yield_df[\"Fertilizer_Used\"].astype(bool)\n",446    "yield_df[\"Irrigation_Used\"] = yield_df[\"Irrigation_Used\"].astype(bool)\n",447    "yield_df[\"Weather_Condition\"] = yield_df[\"Weather_Condition\"].astype(str)\n",448    "yield_df[\"Soil_Type\"] = yield_df[\"Soil_Type\"].astype(str)\n",449    "\n",450    "# 8) Vérification\n",451    "print(\"NaN restants :\")\n",452    "print(yield_df[cols].isna().sum())\n",453    "\n",454    "# 9) Nettoyage optionnel\n",455    "yield_df = yield_df.drop(columns=[\"temp_q\", \"rain_q\"], errors=\"ignore\")\n",456    "\n",457    "yield_df_final = yield_df.copy()"458   ]459  },460  {461   "cell_type": "markdown",462   "id": "b4631f96",463   "metadata": {},464   "source": [465    "### Vérification des distributions"466   ]467  },468  {469   "cell_type": "code",470   "execution_count": 10,471   "id": "db9eb808",472   "metadata": {},473   "outputs": [474    {475     "name": "stdout",476     "output_type": "stream",477     "text": [478      "\n",479      "=== Fertilizer_Used ===\n",480      "Source (crop_yield):\n",481      "Fertilizer_Used\n",482      "False    0.50006\n",483      "True     0.49994\n",484      "Name: proportion, dtype: float64\n",485      "\n",486      "Final (yield_df_final):\n",487      "Fertilizer_Used\n",488      "False    0.501252\n",489      "True     0.498748\n",490      "Name: proportion, dtype: float64\n",491      "\n",492      "=== Irrigation_Used ===\n",493      "Source (crop_yield):\n",494      "Irrigation_Used\n",495      "False    0.500509\n",496      "True     0.499491\n",497      "Name: proportion, dtype: float64\n",498      "\n",499      "Final (yield_df_final):\n",500      "Irrigation_Used\n",501      "True     0.503482\n",502      "False    0.496518\n",503      "Name: proportion, dtype: float64\n",504      "\n",505      "=== Weather_Condition ===\n",506      "Source (crop_yield):\n",507      "Weather_Condition\n",508      "Sunny     0.333790\n",509      "Rainy     0.333561\n",510      "Cloudy    0.332649\n",511      "Name: proportion, dtype: float64\n",512      "\n",513      "Final (yield_df_final):\n",514      "Weather_Condition\n",515      "Cloudy    0.335666\n",516      "Sunny     0.333436\n",517      "Rainy     0.330898\n",518      "Name: proportion, dtype: float64\n",519      "\n",520      "=== Soil_Type ===\n",521      "Source (crop_yield):\n",522      "Soil_Type\n",523      "Sandy     0.167119\n",524      "Loam      0.166795\n",525      "Chalky    0.166779\n",526      "Silt      0.166672\n",527      "Clay      0.166352\n",528      "Peaty     0.166283\n",529      "Name: proportion, dtype: float64\n",530      "\n",531      "Final (yield_df_final):\n",532      "Soil_Type\n",533      "Sandy     0.170286\n",534      "Chalky    0.168159\n",535      "Silt      0.166478\n",536      "Clay      0.165723\n",537      "Peaty     0.165517\n",538      "Loam      0.163837\n",539      "Name: proportion, dtype: float64\n"540     ]541    }542   ],543   "source": [544    "# Pour les variables proxy\n",545    "cols = ['Fertilizer_Used','Irrigation_Used','Weather_Condition','Soil_Type']\n",546    "for col in cols:\n",547    "    print(f\"\\n=== {col} ===\")\n",548    "    print(\"Source (crop_yield):\")\n",549    "    print(crop_yield[col].value_counts(normalize=True))\n",550    "\n",551    "    print(\"\\nFinal (yield_df_final):\")\n",552    "    print(yield_df_final[col].value_counts(normalize=True))"553   ]554  },555  {556   "cell_type": "code",557   "execution_count": 11,558   "id": "84a31ef8",559   "metadata": {},560   "outputs": [561    {562     "name": "stdout",563     "output_type": "stream",564     "text": [565      "\n",566      "=== Fertilizer_Used par crop ===\n",567      "                          source     final\n",568      "       Fertilizer_Used                    \n",569      "Barley False            0.500741       NaN\n",570      "       True             0.499259       NaN\n",571      "Cotton True             0.500609       NaN\n",572      "       False            0.499391       NaN\n",573      "Maize  False            0.501547  0.497063\n",574      "\n",575      "=== Irrigation_Used par crop ===\n",576      "                          source     final\n",577      "       Irrigation_Used                    \n",578      "Barley False            0.500495       NaN\n",579      "       True             0.499505       NaN\n",580      "Cotton False            0.500489       NaN\n",581      "       True             0.499511       NaN\n",582      "Maize  False            0.500042  0.493674\n",583      "\n",584      "=== Weather_Condition par crop ===\n",585      "                            source  final\n",586      "       Weather_Condition                 \n",587      "Barley Rainy              0.334872    NaN\n",588      "       Sunny              0.332954    NaN\n",589      "       Cloudy             0.332174    NaN\n",590      "Cotton Sunny              0.335078    NaN\n",591      "       Rainy              0.333878    NaN\n",592      "\n",593      "=== Soil_Type par crop ===\n",594      "                    source  final\n",595      "       Soil_Type                 \n",596      "Barley Loam       0.167265    NaN\n",597      "       Peaty      0.167031    NaN\n",598      "       Silt       0.166720    NaN\n",599      "       Sandy      0.166396    NaN\n",600      "       Chalky     0.166342    NaN\n"601     ]602    }603   ],604   "source": [605    "# Par type de crop et variable proxy\n",606    "for col in cols:\n",607    "    print(f\"\\n=== {col} par crop ===\")\n",608    "    \n",609    "    source = crop_yield.groupby(\"Crop\")[col].value_counts(normalize=True)\n",610    "    final = yield_df_final.groupby(\"item\")[col].value_counts(normalize=True)\n",611    "    \n",612    "    print(pd.concat([source.rename(\"source\"), final.rename(\"final\")], axis=1).head())"613   ]614  },615  {616   "cell_type": "markdown",617   "id": "8a211fe0",618   "metadata": {},619   "source": [620    "### Séléction des bonnes colonnes"621   ]622  },623  {624   "cell_type": "code",625   "execution_count": 12,626   "id": "e10a916c",627   "metadata": {},628   "outputs": [629    {630     "data": {631      "text/plain": [632       "Index(['Unnamed: 0', 'area', 'region', 'year', 'item', 'avg_temp',\n",633       "       'rainfall_mm', 'pesticides_tonnes', 'yield', 'Fertilizer_Used',\n",634       "       'Irrigation_Used', 'Weather_Condition', 'Soil_Type', 'climate_key'],\n",635       "      dtype='str')"636      ]637     },638     "execution_count": 12,639     "metadata": {},640     "output_type": "execute_result"641    }642   ],643   "source": [644    "yield_df_final.columns"645   ]646  },647  {648   "cell_type": "code",649   "execution_count": 13,650   "id": "9da48f8c",651   "metadata": {},652   "outputs": [],653   "source": [654    "cols = ['area', 'region', 'year', 'item', 'avg_temp', 'rainfall_mm',\n",655    "       'pesticides_tonnes', 'Fertilizer_Used', 'Irrigation_Used',\n",656    "       'Weather_Condition', 'Soil_Type', 'yield']\n",657    "yield_df_enriched = yield_df_final[cols].copy()"658   ]659  },660  {661   "cell_type": "markdown",662   "id": "1e68913b",663   "metadata": {},664   "source": [665    "### Sauvegarde du nouveau fichier"666   ]667  },668  {669   "cell_type": "code",670   "execution_count": 14,671   "id": "3b43c354",672   "metadata": {},673   "outputs": [],674   "source": [675    "yield_df_enriched.to_csv('../data/processed/yield_df_enriched.csv')"676   ]677  },678  {679   "cell_type": "markdown",680   "id": "b2fe173b",681   "metadata": {},682   "source": [683    "## Feature engineering sur le fichier enrichi\n",684    "\n",685    "Nous avons désormais consolidé notre dataset :\n",686    "- Une première fois avec 3 variables climatiques puis imputer de manière rigoureuse pour les données manquantes.\n",687    "- Une deuxième fois en enrichissant le dataset avec des variables approximatives qui devraient aider notre modèle à mieux généraliser\n",688    "\n",689    "Nous devons maintenant continuer l'amélioration de nos données en ajoutant des données si disponibles, en transformant nos données pour être compréhensibles pour un modèle."690   ]691  },692  {693   "cell_type": "markdown",694   "id": "bf237d62",695   "metadata": {},696   "source": [697    "### Nouvelles variables\n",698    "- regroupement des pays : regroupement par grande catégorie pour garder une certaine cohérence\n",699    "- tech_trend : capte la croissance historique des rendements liée aux innovations technologiques au fil des ans.\n",700    "- irrigation_impact : mesure l'importance vitale de l'apport d'eau artificiel, particulièrement élevée lorsque les précipitations naturelles sont insuffisantes.\n",701    "- pest_rain_ratio : évalue l'efficacité potentielle des traitements chimiques en tenant compte du risque de lessivage par les fortes pluies.\n",702    "- climate_instability : identifie les régions à risque en mesurant la variabilité.\n",703    "- relative_tech_intensity : indique si l'effort technologique d'une année spécifique est supérieur ou inférieur à la norme historique du pays concerné."704   ]705  },706  {707   "cell_type": "code",708   "execution_count": 3,709   "id": "653e2fea",710   "metadata": {},711   "outputs": [],712   "source": [713    "yield_df_enriched = pd.read_csv(\"../data/processed/yield_df_enriched.csv\", index_col=0)"714   ]715  },716  {717   "cell_type": "markdown",718   "id": "73b3b880",719   "metadata": {},720   "source": [721    "### Nombre de valeurs par variable"722   ]723  },724  {725   "cell_type": "code",726   "execution_count": 4,727   "id": "e73a2e99",728   "metadata": {},729   "outputs": [730    {731     "data": {732      "text/plain": [733       "item\n",734       "Maize                   4426\n",735       "Potatoes                4163\n",736       "Wheat                   3289\n",737       "Rice                    3171\n",738       "Sweet potatoes          3111\n",739       "Sorghum                 2826\n",740       "Cassava                 2771\n",741       "Soybean                 2428\n",742       "Yams                    1615\n",743       "Plantains and others    1351\n",744       "Name: count, dtype: int64"745      ]746     },747     "execution_count": 4,748     "metadata": {},749     "output_type": "execute_result"750    }751   ],752   "source": [753    "yield_df_enriched['item'].value_counts()"754   ]755  },756  {757   "cell_type": "code",758   "execution_count": 5,759   "id": "b9b2d0e2",760   "metadata": {},761   "outputs": [762    {763     "name": "stdout",764     "output_type": "stream",765     "text": [766      "Fertilizer_Used\n",767      "False    14612\n",768      "True     14539\n",769      "Name: count, dtype: int64\n",770      "==========\n",771      "Irrigation_Used\n",772      "True     14677\n",773      "False    14474\n",774      "Name: count, dtype: int64\n",775      "==========\n",776      "Weather_Condition\n",777      "Cloudy    9785\n",778      "Sunny     9720\n",779      "Rainy     9646\n",780      "Name: count, dtype: int64\n",781      "==========\n",782      "Soil_Type\n",783      "Sandy     4964\n",784      "Chalky    4902\n",785      "Silt      4853\n",786      "Clay      4831\n",787      "Peaty     4825\n",788      "Loam      4776\n",789      "Name: count, dtype: int64\n",790      "==========\n"791     ]792    }793   ],794   "source": [795    "cols = ['Fertilizer_Used', 'Irrigation_Used',\n",796    "       'Weather_Condition', 'Soil_Type']\n",797    "for col in cols:\n",798    "    print(yield_df_enriched[col].value_counts())\n",799    "    print(\"=\"*10)"800   ]801  },802  {803   "cell_type": "markdown",804   "id": "ee398c3e",805   "metadata": {},806   "source": [807    "- On voit que pour les variables catégorielles ci-dessus on va pouvoir utiliser le one hot vu le nombre de valeur limité (colonne supplémentaire limité)"808   ]809  },810  {811   "cell_type": "markdown",812   "id": "9cb5242b",813   "metadata": {},814   "source": [815    "### Ajout des nouvelles variables"816   ]817  },818  {819   "cell_type": "code",820   "execution_count": 6,821   "id": "197604a0",822   "metadata": {},823   "outputs": [],824   "source": [825    "# Ratio climatique - indice d'irrigation critique\n",826    "yield_df_enriched['irrigation_impact'] = yield_df_enriched['Irrigation_Used'].astype(int) / (yield_df_enriched['rainfall_mm'] + 1)\n",827    "# Flag de sécheresse (Binaire) \n",828    "yield_df_enriched['is_drought'] = (yield_df_enriched['rainfall_mm'] < 200).astype(float)\n",829    "# Déséquilibre intrants/eau (Ecart log)\n",830    "yield_df_enriched['input_imbalance'] = np.abs(np.log1p(yield_df_enriched['pesticides_tonnes']) - np.log1p(yield_df_enriched['rainfall_mm']))\n",831    "# Stress thermique simple \n",832    "yield_df_enriched['thermal_stress'] = np.abs(yield_df_enriched['avg_temp'] - 20)\n",833    "# On fixe l'année de référence 2016 pour l'interface future.\n",834    "year_ref = 2016\n",835    "# On crée un score de maturité technologique\n",836    "# Plus l'écart est grand, plus les semences et machines sont supposées performantes.\n",837    "yield_df_enriched['years_from_now'] = year_ref - yield_df_enriched['year']"838   ]839  },840  {841   "cell_type": "code",842   "execution_count": 7,843   "id": "12de3dbd",844   "metadata": {},845   "outputs": [846    {847     "name": "stdout",848     "output_type": "stream",849     "text": [850      "<class 'pandas.core.frame.DataFrame'>\n",851      "Index: 29151 entries, 0 to 29150\n",852      "Data columns (total 17 columns):\n",853      " #   Column             Non-Null Count  Dtype  \n",854      "---  ------             --------------  -----  \n",855      " 0   area               29151 non-null  object \n",856      " 1   region             29151 non-null  object \n",857      " 2   year               29151 non-null  int64  \n",858      " 3   item               29151 non-null  object \n",859      " 4   avg_temp           29151 non-null  float64\n",860      " 5   rainfall_mm        29151 non-null  float64\n",861      " 6   pesticides_tonnes  29151 non-null  float64\n",862      " 7   Fertilizer_Used    29151 non-null  bool   \n",863      " 8   Irrigation_Used    29151 non-null  bool   \n",864      " 9   Weather_Condition  29151 non-null  object \n",865      " 10  Soil_Type          29151 non-null  object \n",866      " 11  yield              29151 non-null  int64  \n",867      " 12  irrigation_impact  29151 non-null  float64\n",868      " 13  is_drought         29151 non-null  float64\n",869      " 14  input_imbalance    29151 non-null  float64\n",870      " 15  thermal_stress     29151 non-null  float64\n",871      " 16  years_from_now     29151 non-null  int64  \n",872      "dtypes: bool(2), float64(7), int64(3), object(5)\n",873      "memory usage: 3.6+ MB\n"874     ]875    }876   ],877   "source": [878    "yield_df_enriched.info()"879   ]880  },881  {882   "cell_type": "markdown",883   "id": "e0209fbe",884   "metadata": {},885   "source": [886    "### Génération des statistiques descriptives"887   ]888  },889  {890   "cell_type": "code",891   "execution_count": 8,892   "id": "b17d227f",893   "metadata": {},894   "outputs": [895    {896     "data": {897      "text/html": [898       "<div>\n",899       "<style scoped>\n",900       "    .dataframe tbody tr th:only-of-type {\n",901       "        vertical-align: middle;\n",902       "    }\n",903       "\n",904       "    .dataframe tbody tr th {\n",905       "        vertical-align: top;\n",906       "    }\n",907       "\n",908       "    .dataframe thead th {\n",909       "        text-align: right;\n",910       "    }\n",911       "</style>\n",912       "<table border=\"1\" class=\"dataframe\">\n",913       "  <thead>\n",914       "    <tr style=\"text-align: right;\">\n",915       "      <th></th>\n",916       "      <th>avg_temp</th>\n",917       "      <th>rainfall_mm</th>\n",918       "      <th>pesticides_tonnes</th>\n",919       "      <th>yield</th>\n",920       "      <th>years_from_now</th>\n",921       "      <th>irrigation_impact</th>\n",922       "      <th>is_drought</th>\n",923       "      <th>input_imbalance</th>\n",924       "      <th>thermal_stress</th>\n",925       "    </tr>\n",926       "  </thead>\n",927       "  <tbody>\n",928       "    <tr>\n",929       "      <th>avg_temp</th>\n",930       "      <td>1.000000</td>\n",931       "      <td>0.369051</td>\n",932       "      <td>-0.328893</td>\n",933       "      <td>-0.077806</td>\n",934       "      <td>-0.002120</td>\n",935       "      <td>-0.105944</td>\n",936       "      <td>0.142590</td>\n",937       "      <td>-0.146340</td>\n",938       "      <td>-0.006201</td>\n",939       "    </tr>\n",940       "    <tr>\n",941       "      <th>rainfall_mm</th>\n",942       "      <td>0.369051</td>\n",943       "      <td>1.000000</td>\n",944       "      <td>-0.016414</td>\n",945       "      <td>0.038608</td>\n",946       "      <td>0.013579</td>\n",947       "      <td>-0.270273</td>\n",948       "      <td>-0.394816</td>\n",949       "      <td>-0.253167</td>\n",950       "      <td>-0.093309</td>\n",951       "    </tr>\n",952       "    <tr>\n",953       "      <th>pesticides_tonnes</th>\n",954       "      <td>-0.328893</td>\n",955       "      <td>-0.016414</td>\n",956       "      <td>1.000000</td>\n",957       "      <td>0.156642</td>\n",958       "      <td>-0.045174</td>\n",959       "      <td>0.008786</td>\n",960       "      <td>-0.105981</td>\n",961       "      <td>0.065299</td>\n",962       "      <td>0.177527</td>\n",963       "    </tr>\n",964       "    <tr>\n",965       "      <th>yield</th>\n",966       "      <td>-0.077806</td>\n",967       "      <td>0.038608</td>\n",968       "      <td>0.156642</td>\n",969       "      <td>1.000000</td>\n",970       "      <td>-0.075102</td>\n",971       "      <td>-0.018039</td>\n",972       "      <td>0.032009</td>\n",973       "      <td>0.025228</td>\n",974       "      <td>0.057594</td>\n",975       "    </tr>\n",976       "    <tr>\n",977       "      <th>years_from_now</th>\n",978       "      <td>-0.002120</td>\n",979       "      <td>0.013579</td>\n",980       "      <td>-0.045174</td>\n",981       "      <td>-0.075102</td>\n",982       "      <td>1.000000</td>\n",983       "      <td>-0.006695</td>\n",984       "      <td>0.004501</td>\n",985       "      <td>0.014067</td>\n",986       "      <td>-0.029976</td>\n",987       "    </tr>\n",988       "    <tr>\n",989       "      <th>irrigation_impact</th>\n",990       "      <td>-0.105944</td>\n",991       "      <td>-0.270273</td>\n",992       "      <td>0.008786</td>\n",993       "      <td>-0.018039</td>\n",994       "      <td>-0.006695</td>\n",995       "      <td>1.000000</td>\n",996       "      <td>0.100936</td>\n",997       "      <td>0.063747</td>\n",998       "      <td>0.022303</td>\n",999       "    </tr>\n",1000       "    <tr>\n",1001       "      <th>is_drought</th>\n",1002       "      <td>0.142590</td>\n",1003       "      <td>-0.394816</td>\n",1004       "      <td>-0.105981</td>\n",1005       "      <td>0.032009</td>\n",1006       "      <td>0.004501</td>\n",1007       "      <td>0.100936</td>\n",1008       "      <td>1.000000</td>\n",1009       "      <td>0.070479</td>\n",1010       "      <td>-0.011864</td>\n",1011       "    </tr>\n",1012       "    <tr>\n",1013       "      <th>input_imbalance</th>\n",1014       "      <td>-0.146340</td>\n",1015       "      <td>-0.253167</td>\n",1016       "      <td>0.065299</td>\n",1017       "      <td>0.025228</td>\n",1018       "      <td>0.014067</td>\n",1019       "      <td>0.063747</td>\n",1020       "      <td>0.070479</td>\n",1021       "      <td>1.000000</td>\n",1022       "      <td>0.062466</td>\n",1023       "    </tr>\n",1024       "    <tr>\n",1025       "      <th>thermal_stress</th>\n",1026       "      <td>-0.006201</td>\n",1027       "      <td>-0.093309</td>\n",1028       "      <td>0.177527</td>\n",1029       "      <td>0.057594</td>\n",1030       "      <td>-0.029976</td>\n",1031       "      <td>0.022303</td>\n",1032       "      <td>-0.011864</td>\n",1033       "      <td>0.062466</td>\n",1034       "      <td>1.000000</td>\n",1035       "    </tr>\n",1036       "  </tbody>\n",1037       "</table>\n",1038       "</div>"1039      ],1040      "text/plain": [1041       "                   avg_temp  rainfall_mm  pesticides_tonnes     yield  \\\n",1042       "avg_temp           1.000000     0.369051          -0.328893 -0.077806   \n",1043       "rainfall_mm        0.369051     1.000000          -0.016414  0.038608   \n",1044       "pesticides_tonnes -0.328893    -0.016414           1.000000  0.156642   \n",1045       "yield             -0.077806     0.038608           0.156642  1.000000   \n",1046       "years_from_now    -0.002120     0.013579          -0.045174 -0.075102   \n",1047       "irrigation_impact -0.105944    -0.270273           0.008786 -0.018039   \n",1048       "is_drought         0.142590    -0.394816          -0.105981  0.032009   \n",1049       "input_imbalance   -0.146340    -0.253167           0.065299  0.025228   \n",1050       "thermal_stress    -0.006201    -0.093309           0.177527  0.057594   \n",1051       "\n",1052       "                   years_from_now  irrigation_impact  is_drought  \\\n",1053       "avg_temp                -0.002120          -0.105944    0.142590   \n",1054       "rainfall_mm              0.013579          -0.270273   -0.394816   \n",1055       "pesticides_tonnes       -0.045174           0.008786   -0.105981   \n",1056       "yield                   -0.075102          -0.018039    0.032009   \n",1057       "years_from_now           1.000000          -0.006695    0.004501   \n",1058       "irrigation_impact       -0.006695           1.000000    0.100936   \n",1059       "is_drought               0.004501           0.100936    1.000000   \n",1060       "input_imbalance          0.014067           0.063747    0.070479   \n",1061       "thermal_stress          -0.029976           0.022303   -0.011864   \n",1062       "\n",1063       "                   input_imbalance  thermal_stress  \n",1064       "avg_temp                 -0.146340       -0.006201  \n",1065       "rainfall_mm              -0.253167       -0.093309  \n",1066       "pesticides_tonnes         0.065299        0.177527  \n",1067       "yield                     0.025228        0.057594  \n",1068       "years_from_now            0.014067       -0.029976  \n",1069       "irrigation_impact         0.063747        0.022303  \n",1070       "is_drought                0.070479       -0.011864  \n",1071       "input_imbalance           1.000000        0.062466  \n",1072       "thermal_stress            0.062466        1.000000  "1073      ]1074     },1075     "execution_count": 8,1076     "metadata": {},1077     "output_type": "execute_result"1078    }1079   ],1080   "source": [1081    "cols_num = [ 'avg_temp', 'rainfall_mm', 'pesticides_tonnes',\n",1082    "            'yield', 'years_from_now', 'irrigation_impact',\n",1083    "            'is_drought','input_imbalance','thermal_stress']\n",1084    "spearman_corr = yield_df_enriched[cols_num].corr(numeric_only=True, method='spearman')\n",1085    "spearman_corr"1086   ]1087  },1088  {1089   "cell_type": "markdown",1090   "id": "f49e77f4",1091   "metadata": {},1092   "source": [1093    "### Matrice des corrélations"1094   ]1095  },1096  {1097   "cell_type": "code",1098   "execution_count": 9,1099   "id": "6a3afc44",1100   "metadata": {},1101   "outputs": [1102    {1103     "data": {1104      "image/png": "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",1105      "text/plain": [1106       "<Figure size 800x600 with 2 Axes>"1107      ]1108     },1109     "metadata": {},1110     "output_type": "display_data"1111    }1112   ],1113   "source": [1114    "plt.figure(figsize=(8, 6))\n",1115    "ax = sns.heatmap(\n",1116    "    spearman_corr,\n",1117    "    cmap=\"coolwarm\",\n",1118    "    center=0,\n",1119    "    annot=True,\n",1120    "    fmt=\".2f\",\n",1121    "    linewidths=0.5)\n",1122    "plt.title(\"Matrice des corrélations - Spearman\", fontsize=16, pad=15)\n",1123    "plt.xticks(rotation=45, ha=\"right\")\n",1124    "plt.yticks(rotation=0)\n",1125    "plt.tight_layout()\n",1126    "plt.savefig(\n",1127    "    \"../notebooks/graph/matrice_correlation_spearman_conso.png\",\n",1128    "    dpi=300,\n",1129    "    bbox_inches=\"tight\",\n",1130    "    facecolor=\"white\")\n",1131    "plt.show()"1132   ]1133  },1134  {1135   "cell_type": "markdown",1136   "id": "ca2d58a4",1137   "metadata": {},1138   "source": [1139    "### Sauvegarde de la dernière mise à jour"1140   ]1141  },1142  {1143   "cell_type": "code",1144   "execution_count": 33,1145   "id": "67dd2678",1146   "metadata": {},1147   "outputs": [],1148   "source": [1149    "df = pd.read_csv(\"../data/processed/yield_df_final_conso_encoded_test.csv\")"1150   ]1151  },1152  {1153   "cell_type": "markdown",1154   "id": "bea4312e",1155   "metadata": {},1156   "source": [1157    "## Feature engineering sur le fichier consolidé\n",1158    "### Chargement du fichier"1159   ]1160  },1161  {1162   "cell_type": "code",1163   "execution_count": 10,1164   "id": "5b2c2291",1165   "metadata": {},1166   "outputs": [],1167   "source": [1168    "yield_df_final = pd.read_csv('../data/processed/yield_df_final.csv', index_col = 0)\n",1169    "yield_df_final_conso = yield_df_final.copy()"1170   ]1171  },1172  {1173   "cell_type": "markdown",1174   "id": "fde17543",1175   "metadata": {},1176   "source": [1177    "### Harmonisation des noms des variables"1178   ]1179  },1180  {1181   "cell_type": "code",1182   "execution_count": 11,1183   "id": "c5798ff9",1184   "metadata": {},1185   "outputs": [1186    {1187     "data": {1188      "text/plain": [1189       "array(['Maize', 'Potatoes', 'Rice', 'Wheat', 'Sorghum', 'Soybean',\n",1190       "       'Cassava', 'Yams', 'Sweet potatoes', 'Plantains and others'],\n",1191       "      dtype=object)"1192      ]1193     },1194     "execution_count": 11,1195     "metadata": {},1196     "output_type": "execute_result"1197    }1198   ],1199   "source": [1200    "yield_df_final_conso['item'] = yield_df_final_conso['item'].replace({\n",

Showing the first 1,200 of 1806 lines. Download the file for the rest.