{
  "id": 480164,
  "title": "How to do EDA?",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/480164",
  "author_name": "Kouhei Miyazaki",
  "post_date": "2024-02-27T13:49:20.424000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I'm relatively new to Kaggle, and excited to join this competition.<br>\nI have a question on how to work on this competition.</p>\n<p>I don't know how to do use findings in EDA. I've checked features in depth = 0 data by using histogram, violin plot, corr() table, … to see the distribution of features, null rate, and so on. </p>\n<p>However, I don't know how to use the finding from it.<br>\nFor example, a lot of features have long-tail shape in histogram. and, have outliers.<br>\nBecause of it, I found the boxplot crushed. So, I checked histogram of these features without outlier and I got a precise distribution of features. </p>\n<p>But,,,what to do next???<br>\nEven though I got it, I can't come up with any idea to get out of this situation.<br>\nIn addition, I don't have knowledge on this field, credit risk.</p>\n<p>I'm glad for any advice. For example, article, notebooks, tips from your experience.<br>\nThank you for your reading =)</p>",
  "messages": [
    {
      "id": 2671414,
      "postDate": "2024-02-27T13:49:20.423Z",
      "content": "<p>I'm relatively new to Kaggle, and excited to join this competition.<br>\nI have a question on how to work on this competition.</p>\n<p>I don't know how to do use findings in EDA. I've checked features in depth = 0 data by using histogram, violin plot, corr() table, … to see the distribution of features, null rate, and so on. </p>\n<p>However, I don't know how to use the finding from it.<br>\nFor example, a lot of features have long-tail shape in histogram. and, have outliers.<br>\nBecause of it, I found the boxplot crushed. So, I checked histogram of these features without outlier and I got a precise distribution of features. </p>\n<p>But,,,what to do next???<br>\nEven though I got it, I can't come up with any idea to get out of this situation.<br>\nIn addition, I don't have knowledge on this field, credit risk.</p>\n<p>I'm glad for any advice. For example, article, notebooks, tips from your experience.<br>\nThank you for your reading =)</p>",
      "rawMarkdown": "I'm relatively new to Kaggle, and excited to join this competition.\nI have a question on how to work on this competition.\n\nI don't know how to do use findings in EDA. I've checked features in depth = 0 data by using histogram, violin plot, corr() table, ... to see the distribution of features, null rate, and so on. \n\nHowever, I don't know how to use the finding from it.\nFor example, a lot of features have long-tail shape in histogram. and, have outliers.\nBecause of it, I found the boxplot crushed. So, I checked histogram of these features without outlier and I got a precise distribution of features. \n\nBut,,,what to do next???\nEven though I got it, I can't come up with any idea to get out of this situation.\nIn addition, I don't have knowledge on this field, credit risk.\n\nI'm glad for any advice. For example, article, notebooks, tips from your experience.\nThank you for your reading =)\n\n",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2671414": "I'm relatively new to Kaggle, and excited to join this competition.\nI have a question on how to work on this competition.\n\nI don't know how to do use findings in EDA. I've checked features in depth = 0 data by using histogram, violin plot, corr() table, ... to see the distribution of features, null rate, and so on. \n\nHowever, I don't know how to use the finding from it.\nFor example, a lot of features have long-tail shape in histogram. and, have outliers.\nBecause of it, I found the boxplot crushed. So, I checked histogram of these features without outlier and I got a precise distribution of features. \n\nBut,,,what to do next???\nEven though I got it, I can't come up with any idea to get out of this situation.\nIn addition, I don't have knowledge on this field, credit risk.\n\nI'm glad for any advice. For example, article, notebooks, tips from your experience.\nThank you for your reading =)\n\n"
  }
}