{
  "id": 475709,
  "title": "Best ways to run the code and handling of null variables",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/475709",
  "author_name": "",
  "post_date": "2024-02-09T13:06:49.127593600Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi people, :)</p>\n<p>I am new to Kaggle challenges and \"real world\" AI applications. So I have two questions and would be super happy if someone could help me. </p>\n<p>First, how are you running your notebooks? Since the datasets are so big I wondered if Google Colab is the best option but I was wondering if something better is out there. </p>\n<p>And second, how are you handling the vast amount of null values in the files? Are you trying to apply data imputation? Or is it better to just use a model that can handle missing data quite well? </p>\n<p>Any help is much appreciated, thanks in advance :) </p>",
  "messages": [
    {
      "id": "2644406",
      "postDate": "02/09/2024 13:06:49",
      "content": "<p>Hi people, :)</p>\n<p>I am new to Kaggle challenges and \"real world\" AI applications. So I have two questions and would be super happy if someone could help me. </p>\n<p>First, how are you running your notebooks? Since the datasets are so big I wondered if Google Colab is the best option but I was wondering if something better is out there. </p>\n<p>And second, how are you handling the vast amount of null values in the files? Are you trying to apply data imputation? Or is it better to just use a model that can handle missing data quite well? </p>\n<p>Any help is much appreciated, thanks in advance :) </p>",
      "rawMarkdown": "Hi people, :)\n\nI am new to Kaggle challenges and \"real world\" AI applications. So I have two questions and would be super happy if someone could help me. \n\nFirst, how are you running your notebooks? Since the datasets are so big I wondered if Google Colab is the best option but I was wondering if something better is out there. \n\nAnd second, how are you handling the vast amount of null values in the files? Are you trying to apply data imputation? Or is it better to just use a model that can handle missing data quite well? \n\nAny help is much appreciated, thanks in advance :)",
      "votes": null
    },
    {
      "id": "2644735",
      "postDate": "02/09/2024 16:32:58",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/matildaknierim\" target=\"_blank\">@matildaknierim</a> </p>\n<p>1 - Google Colab is fine. You can also run notebooks in Kaggle itself. There are many other options as well, <a href=\"https://cloud.jarvislabs.ai/\" target=\"_blank\">jarvislab</a> for example (but its paid).</p>\n<p>2 - Imputation is a very good approach to null values, as it does not waste information. For more detail on that see the Kaggle Course on data cleaning, and machine learning on the learn tab at left.</p>",
      "rawMarkdown": "Hi @matildaknierim \n\n1 - Google Colab is fine. You can also run notebooks in Kaggle itself. There are many other options as well, [jarvislab](https://cloud.jarvislabs.ai/) for example (but its paid).\n\n2 - Imputation is a very good approach to null values, as it does not waste information. For more detail on that see the Kaggle Course on data cleaning, and machine learning on the learn tab at left.",
      "votes": null
    },
    {
      "id": "2646422",
      "postDate": "02/10/2024 21:56:16",
      "content": "<p>Thanks for your help!</p>",
      "rawMarkdown": "Thanks for your help!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2644735,
      "author_name": "gabrielfreddi",
      "author_url": "",
      "post_date": "02/09/2024 16:32:58",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/matildaknierim\" target=\"_blank\">@matildaknierim</a> </p>\n<p>1 - Google Colab is fine. You can also run notebooks in Kaggle itself. There are many other options as well, <a href=\"https://cloud.jarvislabs.ai/\" target=\"_blank\">jarvislab</a> for example (but its paid).</p>\n<p>2 - Imputation is a very good approach to null values, as it does not waste information. For more detail on that see the Kaggle Course on data cleaning, and machine learning on the learn tab at left.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2646422,
          "author_name": "matildaknierim",
          "author_url": "",
          "post_date": "02/10/2024 21:56:16",
          "content": "<p>Thanks for your help!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2644406": "Hi people, :)\n\nI am new to Kaggle challenges and \"real world\" AI applications. So I have two questions and would be super happy if someone could help me. \n\nFirst, how are you running your notebooks? Since the datasets are so big I wondered if Google Colab is the best option but I was wondering if something better is out there. \n\nAnd second, how are you handling the vast amount of null values in the files? Are you trying to apply data imputation? Or is it better to just use a model that can handle missing data quite well? \n\nAny help is much appreciated, thanks in advance :)",
    "2644735": "Hi @matildaknierim \n\n1 - Google Colab is fine. You can also run notebooks in Kaggle itself. There are many other options as well, [jarvislab](https://cloud.jarvislabs.ai/) for example (but its paid).\n\n2 - Imputation is a very good approach to null values, as it does not waste information. For more detail on that see the Kaggle Course on data cleaning, and machine learning on the learn tab at left.",
    "2646422": "Thanks for your help!"
  },
  "source": "meta"
}