{
  "id": 538243,
  "title": "Data Quality Concerns",
  "url": "/competitions/child-mind-institute-problematic-internet-use/discussion/538243",
  "author_name": "",
  "post_date": "2024-10-07T21:30:50.911632500Z",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>As I was exploring the Physical-BMI column lately, I found that there were about 352 samples with a BMI &lt; 15, with some of them having a BMI lower than 10, which, I guess, is extremely low.</p>\n<pre><code>train.loc[train[]&lt;=, [ \n                                     , \n                                     ,  \n                                     ,]]\n</code></pre>\n<p>Also, The Cardiovascular Health Metrics features (Physical-Diastolic_BP, Physical-Systolic_BP, and Physical-HeartRate) exhibit inconsistencies similar to those found in the BMI data.</p>\n<p>I also ensured that Physical-Weight, Physical-Height, and Physical-BMI are completely consistent with the following equation:</p>\n<p>BMI = (Physical-Weight (lbs) * 703) / ( Physical-Height(in))^2</p>\n<p>Any thoughts on how to best handle these values assuming they are data entry errors ?</p>",
  "messages": [
    {
      "id": "3009422",
      "postDate": "10/07/2024 21:30:50",
      "content": "<p>As I was exploring the Physical-BMI column lately, I found that there were about 352 samples with a BMI &lt; 15, with some of them having a BMI lower than 10, which, I guess, is extremely low.</p>\n<pre><code>train.loc[train[]&lt;=, [ \n                                     , \n                                     ,  \n                                     ,]]\n</code></pre>\n<p>Also, The Cardiovascular Health Metrics features (Physical-Diastolic_BP, Physical-Systolic_BP, and Physical-HeartRate) exhibit inconsistencies similar to those found in the BMI data.</p>\n<p>I also ensured that Physical-Weight, Physical-Height, and Physical-BMI are completely consistent with the following equation:</p>\n<p>BMI = (Physical-Weight (lbs) * 703) / ( Physical-Height(in))^2</p>\n<p>Any thoughts on how to best handle these values assuming they are data entry errors ?</p>",
      "rawMarkdown": "As I was exploring the Physical-BMI column lately, I found that there were about 352 samples with a BMI < 15, with some of them having a BMI lower than 10, which, I guess, is extremely low.\n\n```python\ntrain.loc[train[\"Physical-BMI\"]<=15, [\"Physical-Weight\" \n                                     ,\"Physical-Height\" \n                                     , \"Physical-BMI\" \n                                     ,\"Basic_Demos-Age\"]]\n\n```\nAlso, The Cardiovascular Health Metrics features (Physical-Diastolic_BP, Physical-Systolic_BP, and Physical-HeartRate) exhibit inconsistencies similar to those found in the BMI data.\n\nI also ensured that Physical-Weight, Physical-Height, and Physical-BMI are completely consistent with the following equation:\n\nBMI = (Physical-Weight (lbs) * 703) / ( Physical-Height(in))^2\n\nAny thoughts on how to best handle these values assuming they are data entry errors ?",
      "votes": null
    },
    {
      "id": "3009715",
      "postDate": "10/08/2024 08:42:40",
      "content": "<p>There are certainly error entries as some of the BMI is 0. But BMI of &lt;15 isn't as unusual for children as it would be for adults.<br>\nSo you need to take their age into your observation to make a guess if the BMI makes sense.<br>\nHere is a graph where you can see that BMI of 13.5 for 8 years olds is as common as BMI of 18 for 20 years olds, if i understand it correctly.<br>\n<a href=\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1114291/\" target=\"_blank\">NLoM</a></p>",
      "rawMarkdown": "There are certainly error entries as some of the BMI is 0. But BMI of <15 isn't as unusual for children as it would be for adults.\nSo you need to take their age into your observation to make a guess if the BMI makes sense.\nHere is a graph where you can see that BMI of 13.5 for 8 years olds is as common as BMI of 18 for 20 years olds, if i understand it correctly.\n[NLoM](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1114291/)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3009715,
      "author_name": "mariusheuser",
      "author_url": "",
      "post_date": "10/08/2024 08:42:40",
      "content": "<p>There are certainly error entries as some of the BMI is 0. But BMI of &lt;15 isn't as unusual for children as it would be for adults.<br>\nSo you need to take their age into your observation to make a guess if the BMI makes sense.<br>\nHere is a graph where you can see that BMI of 13.5 for 8 years olds is as common as BMI of 18 for 20 years olds, if i understand it correctly.<br>\n<a href=\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1114291/\" target=\"_blank\">NLoM</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3009422": "As I was exploring the Physical-BMI column lately, I found that there were about 352 samples with a BMI < 15, with some of them having a BMI lower than 10, which, I guess, is extremely low.\n\n```python\ntrain.loc[train[\"Physical-BMI\"]<=15, [\"Physical-Weight\" \n                                     ,\"Physical-Height\" \n                                     , \"Physical-BMI\" \n                                     ,\"Basic_Demos-Age\"]]\n\n```\nAlso, The Cardiovascular Health Metrics features (Physical-Diastolic_BP, Physical-Systolic_BP, and Physical-HeartRate) exhibit inconsistencies similar to those found in the BMI data.\n\nI also ensured that Physical-Weight, Physical-Height, and Physical-BMI are completely consistent with the following equation:\n\nBMI = (Physical-Weight (lbs) * 703) / ( Physical-Height(in))^2\n\nAny thoughts on how to best handle these values assuming they are data entry errors ?",
    "3009715": "There are certainly error entries as some of the BMI is 0. But BMI of <15 isn't as unusual for children as it would be for adults.\nSo you need to take their age into your observation to make a guess if the BMI makes sense.\nHere is a graph where you can see that BMI of 13.5 for 8 years olds is as common as BMI of 18 for 20 years olds, if i understand it correctly.\n[NLoM](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1114291/)"
  },
  "source": "meta"
}