{
  "id": 486517,
  "title": "Understanding the Matching of Active and Closed Loans in Bureau Data",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/486517",
  "author_name": "",
  "post_date": "2024-03-25T10:14:07.720953700Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Why, in bureau data, is a borrower's past active or closed loans processed in a form where one active loan is matched with one closed loan?</p>\n<p>In my opinion, this does not seem to be the original format of credit reports, and it is difficult to clearly understand the customer's loan record over time.</p>\n<p>Could the host explain the reason for presenting the data in this way? Is the combination of one active and one closed loan random? Or is there some logic to it?</p>",
  "messages": [
    {
      "id": "2715159",
      "postDate": "03/25/2024 10:14:07",
      "content": "<p>Why, in bureau data, is a borrower's past active or closed loans processed in a form where one active loan is matched with one closed loan?</p>\n<p>In my opinion, this does not seem to be the original format of credit reports, and it is difficult to clearly understand the customer's loan record over time.</p>\n<p>Could the host explain the reason for presenting the data in this way? Is the combination of one active and one closed loan random? Or is there some logic to it?</p>",
      "rawMarkdown": "Why, in bureau data, is a borrower's past active or closed loans processed in a form where one active loan is matched with one closed loan?\n\n In my opinion, this does not seem to be the original format of credit reports, and it is difficult to clearly understand the customer's loan record over time.\n\nCould the host explain the reason for presenting the data in this way? Is the combination of one active and one closed loan random? Or is there some logic to it?",
      "votes": null
    },
    {
      "id": "2715442",
      "postDate": "03/25/2024 13:55:21",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/evan918\" target=\"_blank\">@evan918</a> ,<br>\nI am sorry but I am not sure I understand your question, what do you mean by matching?<br>\nOn our end, we just transformed attributes to \"hide\" information we don't want to or can't share with Kagglers. But there was no matching, or different approach to closed and active loans.</p>",
      "rawMarkdown": "Hi @evan918 ,\nI am sorry but I am not sure I understand your question, what do you mean by matching?\nOn our end, we just transformed attributes to \"hide\" information we don't want to or can't share with Kagglers. But there was no matching, or different approach to closed and active loans.",
      "votes": null
    },
    {
      "id": "2715494",
      "postDate": "03/25/2024 14:27:20",
      "content": "<p>I'm sorry if I wasn't clear earlier.</p>\n<p>For example, in the train_credit_bureau_a_1 dataset, we have many variables corresponding to both active and closed loan, such as dateofcredend_289D (End date of an active credit contract) and dateofcredend_353D (End date of a closed credit contract). This is what I meant by \"matching\".</p>\n<p>In my view, in this dataset one row contains information on two loans, one closed loan and one active loan. If my understanding is incorrect, please correct me.</p>\n<p>I believe the original format of the credit report should be one loan per row, displaying the corresponding status, which might be active or closed, along with the start date and end date. If the status is closed, then there will be a value for the end date; if it's active, it will be empty.</p>",
      "rawMarkdown": "I'm sorry if I wasn't clear earlier.\n\nFor example, in the train_credit_bureau_a_1 dataset, we have many variables corresponding to both active and closed loan, such as dateofcredend_289D (End date of an active credit contract) and dateofcredend_353D (End date of a closed credit contract). This is what I meant by \"matching\".\n\nIn my view, in this dataset one row contains information on two loans, one closed loan and one active loan. If my understanding is incorrect, please correct me.\n\nI believe the original format of the credit report should be one loan per row, displaying the corresponding status, which might be active or closed, along with the start date and end date. If the status is closed, then there will be a value for the end date; if it's active, it will be empty.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2715442,
      "author_name": "tomasjeline2",
      "author_url": "",
      "post_date": "03/25/2024 13:55:21",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/evan918\" target=\"_blank\">@evan918</a> ,<br>\nI am sorry but I am not sure I understand your question, what do you mean by matching?<br>\nOn our end, we just transformed attributes to \"hide\" information we don't want to or can't share with Kagglers. But there was no matching, or different approach to closed and active loans.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2715494,
          "author_name": "evan918",
          "author_url": "",
          "post_date": "03/25/2024 14:27:20",
          "content": "<p>I'm sorry if I wasn't clear earlier.</p>\n<p>For example, in the train_credit_bureau_a_1 dataset, we have many variables corresponding to both active and closed loan, such as dateofcredend_289D (End date of an active credit contract) and dateofcredend_353D (End date of a closed credit contract). This is what I meant by \"matching\".</p>\n<p>In my view, in this dataset one row contains information on two loans, one closed loan and one active loan. If my understanding is incorrect, please correct me.</p>\n<p>I believe the original format of the credit report should be one loan per row, displaying the corresponding status, which might be active or closed, along with the start date and end date. If the status is closed, then there will be a value for the end date; if it's active, it will be empty.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2715159": "Why, in bureau data, is a borrower's past active or closed loans processed in a form where one active loan is matched with one closed loan?\n\n In my opinion, this does not seem to be the original format of credit reports, and it is difficult to clearly understand the customer's loan record over time.\n\nCould the host explain the reason for presenting the data in this way? Is the combination of one active and one closed loan random? Or is there some logic to it?",
    "2715442": "Hi @evan918 ,\nI am sorry but I am not sure I understand your question, what do you mean by matching?\nOn our end, we just transformed attributes to \"hide\" information we don't want to or can't share with Kagglers. But there was no matching, or different approach to closed and active loans.",
    "2715494": "I'm sorry if I wasn't clear earlier.\n\nFor example, in the train_credit_bureau_a_1 dataset, we have many variables corresponding to both active and closed loan, such as dateofcredend_289D (End date of an active credit contract) and dateofcredend_353D (End date of a closed credit contract). This is what I meant by \"matching\".\n\nIn my view, in this dataset one row contains information on two loans, one closed loan and one active loan. If my understanding is incorrect, please correct me.\n\nI believe the original format of the credit report should be one loan per row, displaying the corresponding status, which might be active or closed, along with the start date and end date. If the status is closed, then there will be a value for the end date; if it's active, it will be empty."
  },
  "source": "meta"
}