{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"### American Express - Default Prediction\n\nThe objective of this competition is to predict the probability that a customer does not pay back their credit card balance amount in the future based on their monthly customer profile. The target binary variable is calculated by observing 18 months performance window after the latest credit card statement, and if the customer does not pay due amount in 120 days after their latest statement date it is considered a default event.\n\nThe dataset contains aggregated profile features for each customer at each statement date. Features are anonymized and normalized, and fall into the following general categories:\n\n- D_* = Delinquency variables\n- S_* = Spend variables\n- P_* = Payment variables\n- B_* = Balance variables\n- R_* = Risk variables\n\nwith the following features being categorical:\n\n['B_30', 'B_38', 'D_114', 'D_116', 'D_117', 'D_120', 'D_126', 'D_63', 'D_64', 'D_66', 'D_68']\n\nYour task is to predict, for each customer_ID, the probability of a future payment default (target = 1).\n\nNote that the negative class has been subsampled for this dataset at 5%, and thus receives a 20x weighting in the scoring metric.\n\nThe dataset is too large, so is dificult to read, that's the reason we are using the dataset AMEX-Feather-Dataset from @munum","metadata":{}},{"cell_type":"markdown","source":"## Imports ","metadata":{}},{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport path, os\nfrom pathlib import Path\n\n#Visualization\nimport matplotlib.pyplot as plt\nplt.style.use('bmh')\nimport seaborn as sns\n%matplotlib inline\n\n#Machine Learning\nimport sklearn\nfrom sklearn.model_selection import cross_val_score, RepeatedStratifiedKFold\nfrom sklearn.model_selection import train_test_split, cross_validate, KFold\nfrom sklearn.ensemble import BaggingClassifier\nfrom xgboost import XGBClassifier\nfrom lightgbm import LGBMClassifier\n\nfrom sklearn.metrics import accuracy_score, classification_report, confusion_matrix\n\n\nimport warnings\nwarnings.filterwarnings('ignore')\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-07-07T19:48:44.552088Z","iopub.execute_input":"2022-07-07T19:48:44.552696Z","iopub.status.idle":"2022-07-07T19:48:47.732367Z","shell.execute_reply.started":"2022-07-07T19:48:44.552666Z","shell.execute_reply":"2022-07-07T19:48:47.731048Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Load Data","metadata":{}},{"cell_type":"code","source":"train_data = pd.read_feather('../input/amexfeather/train_data.ftr')\n","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:48:47.735420Z","iopub.execute_input":"2022-07-07T19:48:47.735841Z","iopub.status.idle":"2022-07-07T19:49:08.460066Z","shell.execute_reply.started":"2022-07-07T19:48:47.735796Z","shell.execute_reply":"2022-07-07T19:49:08.459213Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# The testdata will be loaded in other notebook\n#test_data = pd.read_feather('../input/amexfeather/test_data.ftr')","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:08.461330Z","iopub.execute_input":"2022-07-07T19:49:08.461714Z","iopub.status.idle":"2022-07-07T19:49:08.464878Z","shell.execute_reply.started":"2022-07-07T19:49:08.461668Z","shell.execute_reply":"2022-07-07T19:49:08.464111Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## EDA","metadata":{}},{"cell_type":"code","source":"#Training data size\natributes = train_data.shape[0]\ncolumns = train_data.shape[1]\nprint(f'Train data size - Atributes: {atributes}, Columns: {columns}')","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:08.467552Z","iopub.execute_input":"2022-07-07T19:49:08.468097Z","iopub.status.idle":"2022-07-07T19:49:08.482636Z","shell.execute_reply.started":"2022-07-07T19:49:08.468060Z","shell.execute_reply":"2022-07-07T19:49:08.481550Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Train data info\ntrain_data.info(max_cols=200, show_counts=True)","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:08.484024Z","iopub.execute_input":"2022-07-07T19:49:08.484554Z","iopub.status.idle":"2022-07-07T19:49:14.023440Z","shell.execute_reply.started":"2022-07-07T19:49:08.484517Z","shell.execute_reply":"2022-07-07T19:49:14.021630Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"__Insigths__\n\n- There are lots of missing values and type float16, we will decide the best way to handle with this information\n","metadata":{}},{"cell_type":"code","source":"train_data.head()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:14.024721Z","iopub.execute_input":"2022-07-07T19:49:14.025494Z","iopub.status.idle":"2022-07-07T19:49:14.057961Z","shell.execute_reply.started":"2022-07-07T19:49:14.025446Z","shell.execute_reply":"2022-07-07T19:49:14.056914Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Checking the null values\n\ntrain_data.isnull().sum()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:14.059296Z","iopub.execute_input":"2022-07-07T19:49:14.059760Z","iopub.status.idle":"2022-07-07T19:49:19.395508Z","shell.execute_reply.started":"2022-07-07T19:49:14.059722Z","shell.execute_reply":"2022-07-07T19:49:19.394723Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- Column D_142 has lots of null values, maybe could be dropped","metadata":{}},{"cell_type":"code","source":"#checking the null values in percent\n\nnull_values = round(train_data.isnull().sum()/train_data.shape[0]*100,2).sort_values(ascending=False)","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:19.396961Z","iopub.execute_input":"2022-07-07T19:49:19.397419Z","iopub.status.idle":"2022-07-07T19:49:24.345849Z","shell.execute_reply.started":"2022-07-07T19:49:19.397381Z","shell.execute_reply":"2022-07-07T19:49:24.344977Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"null_values_df = null_values.to_frame(name='Missing Values %')\nnull_values_df.head(10)","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:24.347308Z","iopub.execute_input":"2022-07-07T19:49:24.347717Z","iopub.status.idle":"2022-07-07T19:49:24.358100Z","shell.execute_reply.started":"2022-07-07T19:49:24.347641Z","shell.execute_reply":"2022-07-07T19:49:24.357041Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"__Insigths__\n\n- We have in this dataset some columns that has almost 100% of missing values.\n- I will check which those columns has more than 50 % and 90% of missing and later We will decide how can we handle with that.","metadata":{}},{"cell_type":"code","source":"# Greater than 50% of missing values\nnull_granter_than_50 = null_values_df[null_values_df['Missing Values %'] >= 50.00]\nprint(f'Number of columns with more than 50% of missing values: {null_granter_than_50.shape[0]}')\nnull_granter_than_50.head(10)","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:24.361773Z","iopub.execute_input":"2022-07-07T19:49:24.362322Z","iopub.status.idle":"2022-07-07T19:49:24.377608Z","shell.execute_reply.started":"2022-07-07T19:49:24.362284Z","shell.execute_reply":"2022-07-07T19:49:24.376852Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Greate than 90% og missing values\nnull_granter_than_90 = null_values_df[null_values_df['Missing Values %'] >= 90.00]\nprint(f'Number of columns with more than 50% of missing values: {null_granter_than_90.shape[0]}')\nnull_granter_than_90.head(18)","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:24.379021Z","iopub.execute_input":"2022-07-07T19:49:24.379937Z","iopub.status.idle":"2022-07-07T19:49:24.394741Z","shell.execute_reply.started":"2022-07-07T19:49:24.379900Z","shell.execute_reply":"2022-07-07T19:49:24.393909Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Statistical information\ntrain_data.describe()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:49:24.396639Z","iopub.execute_input":"2022-07-07T19:49:24.397591Z","iopub.status.idle":"2022-07-07T19:51:56.866090Z","shell.execute_reply.started":"2022-07-07T19:49:24.397551Z","shell.execute_reply":"2022-07-07T19:51:56.865278Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- check de describe again after pre processing data to analyze better","metadata":{}},{"cell_type":"code","source":"#Target Distribution to check imbalanced class\n\nprint(f\"Classe Target percents: \\n{round(train_data['target'].value_counts(normalize=True),2)}\")","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:51:56.867245Z","iopub.execute_input":"2022-07-07T19:51:56.867813Z","iopub.status.idle":"2022-07-07T19:51:56.900302Z","shell.execute_reply.started":"2022-07-07T19:51:56.867776Z","shell.execute_reply":"2022-07-07T19:51:56.899418Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize=(8,5))\nsns.countplot(x='target', data=train_data, palette='Blues')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:54:42.926088Z","iopub.execute_input":"2022-07-07T19:54:42.926463Z","iopub.status.idle":"2022-07-07T19:54:43.531506Z","shell.execute_reply.started":"2022-07-07T19:54:42.926435Z","shell.execute_reply":"2022-07-07T19:54:43.530657Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\"The training data begins on {} and ends on {}.\".format(train_data['S_2'].min().strftime('%m-%d-%Y'),train_data['S_2'].max().strftime('%m-%d-%Y')))","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:51:57.563160Z","iopub.execute_input":"2022-07-07T19:51:57.563681Z","iopub.status.idle":"2022-07-07T19:51:57.603266Z","shell.execute_reply.started":"2022-07-07T19:51:57.563644Z","shell.execute_reply":"2022-07-07T19:51:57.602297Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Check customer duplicated\ntrain_data['customer_ID'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:51:57.604617Z","iopub.execute_input":"2022-07-07T19:51:57.605574Z","iopub.status.idle":"2022-07-07T19:51:58.563611Z","shell.execute_reply.started":"2022-07-07T19:51:57.605532Z","shell.execute_reply":"2022-07-07T19:51:58.562768Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#analysing the number of statement of customer default\ncustomer_default = train_data[train_data['target'] == 1]","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:51:58.565084Z","iopub.execute_input":"2022-07-07T19:51:58.565517Z","iopub.status.idle":"2022-07-07T19:52:00.807222Z","shell.execute_reply.started":"2022-07-07T19:51:58.565478Z","shell.execute_reply":"2022-07-07T19:52:00.806341Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\ncustomer_default['customer_ID'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:00.808634Z","iopub.execute_input":"2022-07-07T19:52:00.809017Z","iopub.status.idle":"2022-07-07T19:52:01.035993Z","shell.execute_reply.started":"2022-07-07T19:52:00.808980Z","shell.execute_reply":"2022-07-07T19:52:01.035000Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Check the statements information from one customer\ntrain_data[train_data['customer_ID'] == '0000f99513770170a1aba690daeeb8a96da4a39f11fc27da5c30a79db61c1e85']","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:01.037491Z","iopub.execute_input":"2022-07-07T19:52:01.037924Z","iopub.status.idle":"2022-07-07T19:52:01.843657Z","shell.execute_reply.started":"2022-07-07T19:52:01.037889Z","shell.execute_reply":"2022-07-07T19:52:01.842829Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"__Insigths__\n\n- Its not a duplicated values, its a statements per costumer\n- We see that this exemple of customer default has 1 statements per months for a year","metadata":{}},{"cell_type":"code","source":"#Analysing the statements x target\nplt.figure(figsize=(10, 6))\nsns.countplot(x=train_data.groupby(\"customer_ID\")['target'].max(), palette='Blues')\nplt.title(\"Distribution of the number of statements x target\", fontsize=14)\nplt.xlabel(\"Statements\", fontsize=12)\nplt.ylabel(\"Target\", fontsize=12);","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:01.845047Z","iopub.execute_input":"2022-07-07T19:52:01.845419Z","iopub.status.idle":"2022-07-07T19:52:03.497479Z","shell.execute_reply.started":"2022-07-07T19:52:01.845383Z","shell.execute_reply":"2022-07-07T19:52:03.496700Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- The most of customer who has more stetements are not defalt","metadata":{}},{"cell_type":"markdown","source":"## Categorical features\n","metadata":{"execution":{"iopub.status.busy":"2022-07-04T01:29:40.609762Z","iopub.execute_input":"2022-07-04T01:29:40.610431Z","iopub.status.idle":"2022-07-04T01:29:40.614Z","shell.execute_reply.started":"2022-07-04T01:29:40.610395Z","shell.execute_reply":"2022-07-04T01:29:40.613225Z"}}},{"cell_type":"code","source":"categorical_features = ['B_30', 'B_38', 'D_114', 'D_116', 'D_117', 'D_120', 'D_126', \n                            'D_63', 'D_64', 'D_66', 'D_68']\n\nplt.figure(figsize=(15,9))\nfor i in range(0,len(categorical_features)):\n    plt.subplot(4,3, i + 1)\n    sns.countplot(data= train_data, x = categorical_features[i], hue='target', palette='tab10')\n    plt.tight_layout()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:03.498931Z","iopub.execute_input":"2022-07-07T19:52:03.499330Z","iopub.status.idle":"2022-07-07T19:52:10.421970Z","shell.execute_reply.started":"2022-07-07T19:52:03.499292Z","shell.execute_reply":"2022-07-07T19:52:10.421142Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"__Insigths__\n\n- The most of feature has more than 2 category, so encode can fit.\n- Features with more correlation with target = B_38, D_117, D_126, D_63, D_64, D_68\n- The feature that will not be droppred, we can do one hot encode","metadata":{}},{"cell_type":"markdown","source":"## Numerical Features","metadata":{}},{"cell_type":"code","source":"features_non_numerical = categorical_features + ['customer_ID', 'S_2']\n\nnumerical_features = [i for i in train_data.columns if i not in features_non_numerical ]\nprint(f'Number of numerical features: {len(numerical_features)}')\n","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:10.426326Z","iopub.execute_input":"2022-07-07T19:52:10.428527Z","iopub.status.idle":"2022-07-07T19:52:10.438875Z","shell.execute_reply.started":"2022-07-07T19:52:10.428488Z","shell.execute_reply":"2022-07-07T19:52:10.438094Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"d_features = list(filter(lambda x: x.startswith(\"D_\"), numerical_features)) + ['target']\ns_features = list(filter(lambda x: x.startswith(\"S_\"), numerical_features)) + ['target']\np_features = list(filter(lambda x: x.startswith(\"P_\"), numerical_features)) + ['target']\nb_features = list(filter(lambda x: x.startswith(\"B_\"), numerical_features)) + ['target']\nr_features = list(filter(lambda x: x.startswith(\"R_\"), numerical_features)) + ['target']","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:10.441014Z","iopub.execute_input":"2022-07-07T19:52:10.443013Z","iopub.status.idle":"2022-07-07T19:52:10.456083Z","shell.execute_reply.started":"2022-07-07T19:52:10.442976Z","shell.execute_reply":"2022-07-07T19:52:10.455373Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Correlation Delinquency Variables\nd_features_corr = train_data[d_features].corr()\nd_features_corr","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:52:10.460360Z","iopub.execute_input":"2022-07-07T19:52:10.461831Z","iopub.status.idle":"2022-07-07T19:53:34.088895Z","shell.execute_reply.started":"2022-07-07T19:52:10.461794Z","shell.execute_reply":"2022-07-07T19:53:34.087879Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Correlation Spend Features\ns_features_corr = train_data[s_features].corr()\nS_features_corr_copy = s_features_corr.iloc[1:, :-1]\nmask=np.triu(np.ones_like(s_features_corr, dtype=bool))[1:,:-1]\nplt.figure(figsize=(16,12))\nsns.heatmap(S_features_corr_copy,mask=mask, cmap='Blues',vmin=-1, vmax=1, center=0, annot=True, fmt='.2f', annot_kws={'fontsize':10,'fontweight':'bold'}, cbar=False)\nplt.title(\"Correlation Spend Variables\")\nplt.tight_layout()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:53:34.090380Z","iopub.execute_input":"2022-07-07T19:53:34.090822Z","iopub.status.idle":"2022-07-07T19:53:43.496388Z","shell.execute_reply.started":"2022-07-07T19:53:34.090784Z","shell.execute_reply":"2022-07-07T19:53:43.495517Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- Most positive correlation with Target = S_3, S_7, S_15\n- Most negative correlation wit target = S_25, S_8","metadata":{}},{"cell_type":"code","source":"#Correlation Payment Features\np_features_corr = train_data[p_features].corr()\np_features_corr_copy = p_features_corr.iloc[1:, :-1]\nmask=np.triu(np.ones_like(p_features_corr, dtype=bool))[1:,:-1]\nplt.figure(figsize=(8,6))\nsns.heatmap(p_features_corr_copy,mask=mask, cmap='Blues',vmin=-1, vmax=1, center=0, annot=True, fmt='.2f', annot_kws={'fontsize':10,'fontweight':'bold'}, cbar=False)\nplt.title(\"Correlation Payment Variables\")\nplt.tight_layout()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:53:43.497348Z","iopub.execute_input":"2022-07-07T19:53:43.497714Z","iopub.status.idle":"2022-07-07T19:53:44.131221Z","shell.execute_reply.started":"2022-07-07T19:53:43.497683Z","shell.execute_reply":"2022-07-07T19:53:44.130227Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- Positive correlation with target = P_4\n- Negative correlation = P_2, P_3","metadata":{}},{"cell_type":"code","source":"#Correlation Balance Features\n\nb_features_corr = train_data[b_features].corr()\nb_features_corr_copy = b_features_corr.iloc[1:, :-1]\nmask=np.triu(np.ones_like(b_features_corr, dtype=bool))[1:,:-1]\nplt.figure(figsize=(18,16))\nsns.heatmap(b_features_corr_copy,mask=mask, cmap='Blues',vmin=-1, vmax=1, center=0, annot=True, fmt='.2f', annot_kws={'fontsize':10,'fontweight':'bold'}, cbar=False)\nplt.title(\"Correlation Balance Variables\")\nplt.tight_layout()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:53:44.136134Z","iopub.execute_input":"2022-07-07T19:53:44.136751Z","iopub.status.idle":"2022-07-07T19:54:08.870626Z","shell.execute_reply.started":"2022-07-07T19:53:44.136704Z","shell.execute_reply":"2022-07-07T19:54:08.869671Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- Some Positive correlation = B_1, B_7, B_8, B_9, B_11, B_16\n- Some Negative correlation = B_2, B_18","metadata":{}},{"cell_type":"code","source":"#Correlation Risk Features\n\nr_features_corr = train_data[r_features].corr()\nr_features_corr_copy = r_features_corr.iloc[1:, :-1]\nmask=np.triu(np.ones_like(r_features_corr, dtype=bool))[1:,:-1]\nplt.figure(figsize=(18,16))\nsns.heatmap(r_features_corr_copy,mask=mask, cmap='Blues',vmin=-1, vmax=1, center=0, annot=True, fmt='.2f', annot_kws={'fontsize':10,'fontweight':'bold'}, cbar=False)\nplt.title(\"Correlation Risk Variables\")\nplt.tight_layout()","metadata":{"execution":{"iopub.status.busy":"2022-07-07T19:54:08.874462Z","iopub.execute_input":"2022-07-07T19:54:08.874819Z","iopub.status.idle":"2022-07-07T19:54:24.369006Z","shell.execute_reply.started":"2022-07-07T19:54:08.874788Z","shell.execute_reply":"2022-07-07T19:54:24.368151Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Analysis: \n\n- Handle with missing values - some colunms has almost 100% of missing values\n- The most columns that has more than 90% of missing values are Delinquency\n- This dataset is imbalanced, its is 25% target default\n- we have one statements a months for customer for one year, the most statement are non default\n-  Thare are 6 categorical feature with more correlation, and more than 2 category","metadata":{}}]}