hi dear kagglers:
\nI'm new to this compitation and still doing EDA. I found something in table: credit_bureau_a which helps me to gain a deeper understanding of this dataset. And I want to share it: I guess the organizer concatenated two tables into one and provided it to us. One table is for active contracts and the other table is for closed contracts. We should also analyze the data separately.
\ntrain_credit_bureau_a_1_all = pd.DataFrame()\n\n i tqdm(()):\n ()\n train_credit_bureau_a_1_X = pd.read_csv(join(csv_train_path, ))\n case_id_temp = train_credit_bureau_a_1_X[].unique().tolist()\n\n train_credit_bureau_a_1_all = pd.concat(\n [train_credit_bureau_a_1_all, train_credit_bureau_a_1_X], axis=)\n\n\ntrain_credit_bureau_a_1_all.loc[train_credit_bureau_a_1_all[]==][].unique()\n\n\n\n\n\nAm I right or not, what's your opinion?😀
", "messages": [ { "id": 2753089, "postDate": "2024-04-15T09:32:02.360Z", "content": "hi dear kagglers:
\nI'm new to this compitation and still doing EDA. I found something in table: credit_bureau_a which helps me to gain a deeper understanding of this dataset. And I want to share it: I guess the organizer concatenated two tables into one and provided it to us. One table is for active contracts and the other table is for closed contracts. We should also analyze the data separately.
\ntrain_credit_bureau_a_1_all = pd.DataFrame()\n\n i tqdm(()):\n ()\n train_credit_bureau_a_1_X = pd.read_csv(join(csv_train_path, ))\n case_id_temp = train_credit_bureau_a_1_X[].unique().tolist()\n\n train_credit_bureau_a_1_all = pd.concat(\n [train_credit_bureau_a_1_all, train_credit_bureau_a_1_X], axis=)\n\n\ntrain_credit_bureau_a_1_all.loc[train_credit_bureau_a_1_all[]==][].unique()\n\n\n\n\n\nAm I right or not, what's your opinion?😀
", "rawMarkdown": "hi dear kagglers:\n \nI'm new to this compitation and still doing EDA. I found something in table: credit_bureau_a which helps me to gain a deeper understanding of this dataset. And I want to share it: **I guess the organizer concatenated two tables into one and provided it to us. One table is for active contracts and the other table is for closed contracts. We should also analyze the data separately.**\n\n \n```python\n\ntrain_credit_bureau_a_1_all = pd.DataFrame()\n\nfor i in tqdm(range(4)):\n print(f'**************** {i} ***************')\n train_credit_bureau_a_1_X = pd.read_csv(join(csv_train_path, f'train_credit_bureau_a_1_{i}.csv'))\n case_id_temp = train_credit_bureau_a_1_X['case_id'].unique().tolist()\n \n train_credit_bureau_a_1_all = pd.concat(\n [train_credit_bureau_a_1_all, train_credit_bureau_a_1_X], axis=0)\n\n\ntrain_credit_bureau_a_1_all.loc[train_credit_bureau_a_1_all['financialinstitution_591M']=='a55475b1']['dateofcredend_289D'].unique()\n\n# financialinstitution_591M: Financial institution name of the active contract.\n# dateofcredend_289D: End date of an active credit contract.\n\n# you will find the result is: array([nan], dtype=object), and I guass it's the rule to select real active credit contract and 'a55475b1' means 'unknow'?\n\n```\n\nAm I right or not, what's your opinion?😀\n\n\n\n\n\n\n\n\n\n", "votes": 13 }, { "id": 2753203, "postDate": "2024-04-15T11:05:24.800Z", "content": "that's a great way to look at it!
", "rawMarkdown": "that's a great way to look at it!", "votes": 1 }, { "id": 2753390, "postDate": "2024-04-15T13:16:07.663Z", "content": "I had a similar thought and I asked how active and closed loans are matched in another discussion, but didn't get an answer. https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/486517#2715494
", "rawMarkdown": "I had a similar thought and I asked how active and closed loans are matched in another discussion, but didn't get an answer. https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/486517#2715494", "votes": 2, "replies": [ { "id": 2753487, "postDate": "2024-04-15T14:07:58.493Z", "content": "The process of understanding this dataset is a bit like solving a puzzle.😅
", "rawMarkdown": "The process of understanding this dataset is a bit like solving a puzzle.😅", "votes": 1, "replies": [ { "id": 2753540, "postDate": "2024-04-15T14:54:33.513Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ] } ] }, { "id": 2756098, "postDate": "2024-04-16T21:21:53.747Z", "content": "Yes,right
", "rawMarkdown": "Yes,right " }, { "id": 2753090, "postDate": "2024-04-15T09:33:19.513Z", "content": "If the understanding is right, it may help us to create feature more appropriately🥳
", "rawMarkdown": "If the understanding is right, it may help us to create feature more appropriately🥳" }, { "id": 2767523, "postDate": "2024-04-22T11:19:54.247Z", "content": "Great insight, thank you!
", "rawMarkdown": "Great insight, thank you!" } ], "comments": [ { "id": 2753203, "author_name": "Akul Vaishnavi", "author_url": "", "post_date": "2024-04-15T11:05:24.800000", "content": "that's a great way to look at it!
", "votes": 1, "replies": [] }, { "id": 2753390, "author_name": "Evan", "author_url": "", "post_date": "2024-04-15T13:16:07.663000", "content": "I had a similar thought and I asked how active and closed loans are matched in another discussion, but didn't get an answer. https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/486517#2715494
", "votes": 2, "replies": [ { "id": 2753487, "author_name": "RogerOcean", "author_url": "", "post_date": "2024-04-15T14:07:58.493000", "content": "The process of understanding this dataset is a bit like solving a puzzle.😅
", "votes": 1, "replies": [ { "id": 2753540, "author_name": "", "author_url": "", "post_date": "2024-04-15T14:54:33.513000", "content": "", "votes": -1, "replies": [] } ] } ] }, { "id": 2756098, "author_name": "", "author_url": "", "post_date": "2024-04-16T21:21:53.747000", "content": "Yes,right
", "votes": 0, "replies": [] }, { "id": 2753090, "author_name": "RogerOcean", "author_url": "", "post_date": "2024-04-15T09:33:19.513000", "content": "If the understanding is right, it may help us to create feature more appropriately🥳
", "votes": 0, "replies": [] }, { "id": 2767523, "author_name": "Cony Kuo", "author_url": "", "post_date": "2024-04-22T11:19:54.247000", "content": "Great insight, thank you!
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "2753089": "hi dear kagglers:\n \nI'm new to this compitation and still doing EDA. I found something in table: credit_bureau_a which helps me to gain a deeper understanding of this dataset. And I want to share it: **I guess the organizer concatenated two tables into one and provided it to us. One table is for active contracts and the other table is for closed contracts. We should also analyze the data separately.**\n\n \n```python\n\ntrain_credit_bureau_a_1_all = pd.DataFrame()\n\nfor i in tqdm(range(4)):\n print(f'**************** {i} ***************')\n train_credit_bureau_a_1_X = pd.read_csv(join(csv_train_path, f'train_credit_bureau_a_1_{i}.csv'))\n case_id_temp = train_credit_bureau_a_1_X['case_id'].unique().tolist()\n \n train_credit_bureau_a_1_all = pd.concat(\n [train_credit_bureau_a_1_all, train_credit_bureau_a_1_X], axis=0)\n\n\ntrain_credit_bureau_a_1_all.loc[train_credit_bureau_a_1_all['financialinstitution_591M']=='a55475b1']['dateofcredend_289D'].unique()\n\n# financialinstitution_591M: Financial institution name of the active contract.\n# dateofcredend_289D: End date of an active credit contract.\n\n# you will find the result is: array([nan], dtype=object), and I guass it's the rule to select real active credit contract and 'a55475b1' means 'unknow'?\n\n```\n\nAm I right or not, what's your opinion?😀\n\n\n\n\n\n\n\n\n\n", "2753203": "that's a great way to look at it!", "2753390": "I had a similar thought and I asked how active and closed loans are matched in another discussion, but didn't get an answer. https://www.kaggle.com/competitions/home-credit-credit-risk-model-stability/discussion/486517#2715494", "2756098": "Yes,right ", "2753090": "If the understanding is right, it may help us to create feature more appropriately🥳", "2767523": "Great insight, thank you!" } }