{
  "id": 285743,
  "title": " Where do you get code samples from?",
  "url": "/competitions/g-research-crypto-forecasting/discussion/285743",
  "author_name": "",
  "post_date": "2021-11-06T03:35:02.952299Z",
  "votes": 6,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I know Kaggle is a great plateform for ML, but I wonder how people are getting deep learning sample code for for their competitions. </p>\n<p>For example I need sample code for library 1 but many time it is ending in error or due to outdated keyword. I know those baseline code is built using someother dataset, but if I alter the code for my dataset its getting error in most of the case. Can I know  </p>\n<ol>\n<li><p>How you are Getting the deep learning sample code ?<br>\n1.1. In Kaggle we have plenty of code but is their any platform apart from kaggle like github repo, etc..</p></li>\n<li><p>Are you memorizing the library's keyword  and building the code from scratch?</p></li>\n<li><p>Are you modifying the existing code?. </p></li>\n</ol>\n<p>If I am anything wrong please apologize me. Thanks in advance</p>",
  "messages": [
    {
      "id": "1572868",
      "postDate": "11/06/2021 03:35:02",
      "content": "<p>I know Kaggle is a great plateform for ML, but I wonder how people are getting deep learning sample code for for their competitions. </p>\n<p>For example I need sample code for library 1 but many time it is ending in error or due to outdated keyword. I know those baseline code is built using someother dataset, but if I alter the code for my dataset its getting error in most of the case. Can I know  </p>\n<ol>\n<li><p>How you are Getting the deep learning sample code ?<br>\n1.1. In Kaggle we have plenty of code but is their any platform apart from kaggle like github repo, etc..</p></li>\n<li><p>Are you memorizing the library's keyword  and building the code from scratch?</p></li>\n<li><p>Are you modifying the existing code?. </p></li>\n</ol>\n<p>If I am anything wrong please apologize me. Thanks in advance</p>",
      "rawMarkdown": "I know Kaggle is a great plateform for ML, but I wonder how people are getting deep learning sample code for for their competitions. \n\nFor example I need sample code for library 1 but many time it is ending in error or due to outdated keyword. I know those baseline code is built using someother dataset, but if I alter the code for my dataset its getting error in most of the case. Can I know  \n\n1. How you are Getting the deep learning sample code ?\n1.1. In Kaggle we have plenty of code but is their any platform apart from kaggle like github repo, etc..\n\n2. Are you memorizing the library's keyword  and building the code from scratch?\n\n3. Are you modifying the existing code?. \n\n\nIf I am anything wrong please apologize me. Thanks in advance",
      "votes": null
    },
    {
      "id": "1573084",
      "postDate": "11/06/2021 08:54:18",
      "content": "<p>Why do you need to memorize ? Normally people look up documentation when they write code. Most of the library comes with one.</p>",
      "rawMarkdown": "Why do you need to memorize ? Normally people look up documentation when they write code. Most of the library comes with one.",
      "votes": null
    },
    {
      "id": "1573167",
      "postDate": "11/06/2021 11:18:27",
      "content": "<hr>\n<p><strong>I just want to point something important:</strong> The reason people downvoted this post is not because of the question itself. <br>\nThis is mainly because of the culture of Kaggle. [This is my own opinion that might be wrong] </p>\n<p>People downvote this post despite it containing a legitimate question simply because the title doesn't detail the question itself ['Help! Help! Help!']. I get it. you seek some help from the community and there are many different communities where the title 'Help! Help! Help!' will be accepted.  I think that editing it to something like: \"Help! Help! Help! Where do you get code examples from?\" will encourage people not only to upvote it but to even come and help. since it still has your \"personality in it 'Help! Help! Help!' but also it fits the kaggle culture of meaningful posts.   </p>\n<hr>\n<p><strong>To the question:</strong> By far the best method to get code examples for kaggle is.. well.. from kaggle. If you can't find what you look for in this competition (since it is new.. it takes time) you can always go back to previous related competitions. Often times you can find a list detailing past relevant competitions on the discussions. Just pay them a visit. You will find plenty of working end2end code on them. This is by far the best method for getting new ideas and pieces of code for kaggle. [In my opinion: Often for real-world machine learning projects as well] </p>\n<hr>\n<p>Good Luck and don't be discouraged by the downvoters.<br>\nThey have nothing personal against you</p>",
      "rawMarkdown": "____\n\n**I just want to point something important:** The reason people downvoted this post is not because of the question itself. \nThis is mainly because of the culture of Kaggle. [This is my own opinion that might be wrong] \n\nPeople downvote this post despite it containing a legitimate question simply because the title doesn't detail the question itself ['Help! Help! Help!']. I get it. you seek some help from the community and there are many different communities where the title 'Help! Help! Help!' will be accepted.  I think that editing it to something like: \"Help! Help! Help! Where do you get code examples from?\" will encourage people not only to upvote it but to even come and help. since it still has your \"personality in it 'Help! Help! Help!' but also it fits the kaggle culture of meaningful posts.   \n\n____\n\n**To the question:** By far the best method to get code examples for kaggle is.. well.. from kaggle. If you can't find what you look for in this competition (since it is new.. it takes time) you can always go back to previous related competitions. Often times you can find a list detailing past relevant competitions on the discussions. Just pay them a visit. You will find plenty of working end2end code on them. This is by far the best method for getting new ideas and pieces of code for kaggle. [In my opinion: Often for real-world machine learning projects as well] \n\n____\n\nGood Luck and don't be discouraged by the downvoters.\nThey have nothing personal against you",
      "votes": null
    },
    {
      "id": "1573185",
      "postDate": "11/06/2021 11:34:06",
      "content": "<p>I have corrected the heading after your comment. Thank you so much for your response and taking time to comment </p>",
      "rawMarkdown": "I have corrected the heading after your comment. Thank you so much for your response and taking time to comment",
      "votes": null
    },
    {
      "id": "1573190",
      "postDate": "11/06/2021 11:39:27",
      "content": "<p>the official documentation of any library is enough, for example, if you want to use any functions built under pandas library you can check the documentation of that function on Pandas official website and you'll find an explication of all arguments + some code examples of using it (for ex the usage of <a href=\"https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.groupby.html\" target=\"_blank\">groupby</a>), and the same with any other library like seaborn, scikit learn ,numpy …</p>",
      "rawMarkdown": "the official documentation of any library is enough, for example, if you want to use any functions built under pandas library you can check the documentation of that function on Pandas official website and you'll find an explication of all arguments + some code examples of using it (for ex the usage of [groupby](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.groupby.html)), and the same with any other library like seaborn, scikit learn ,numpy ...",
      "votes": null
    },
    {
      "id": "1615430",
      "postDate": "12/12/2021 08:34:09",
      "content": "<p>I learned python and machine learning from MOOC. In those courses, there are plenty of case studies, some are quite similar to Kaggle competiton's problem statement. We just need to learning by doing.;)</p>",
      "rawMarkdown": "I learned python and machine learning from MOOC. In those courses, there are plenty of case studies, some are quite similar to Kaggle competiton's problem statement. We just need to learning by doing.;)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1573084,
      "author_name": "zarif98sjs",
      "author_url": "",
      "post_date": "11/06/2021 08:54:18",
      "content": "<p>Why do you need to memorize ? Normally people look up documentation when they write code. Most of the library comes with one.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1573167,
      "author_name": "",
      "author_url": "",
      "post_date": "11/06/2021 11:18:27",
      "content": "<hr>\n<p><strong>I just want to point something important:</strong> The reason people downvoted this post is not because of the question itself. <br>\nThis is mainly because of the culture of Kaggle. [This is my own opinion that might be wrong] </p>\n<p>People downvote this post despite it containing a legitimate question simply because the title doesn't detail the question itself ['Help! Help! Help!']. I get it. you seek some help from the community and there are many different communities where the title 'Help! Help! Help!' will be accepted.  I think that editing it to something like: \"Help! Help! Help! Where do you get code examples from?\" will encourage people not only to upvote it but to even come and help. since it still has your \"personality in it 'Help! Help! Help!' but also it fits the kaggle culture of meaningful posts.   </p>\n<hr>\n<p><strong>To the question:</strong> By far the best method to get code examples for kaggle is.. well.. from kaggle. If you can't find what you look for in this competition (since it is new.. it takes time) you can always go back to previous related competitions. Often times you can find a list detailing past relevant competitions on the discussions. Just pay them a visit. You will find plenty of working end2end code on them. This is by far the best method for getting new ideas and pieces of code for kaggle. [In my opinion: Often for real-world machine learning projects as well] </p>\n<hr>\n<p>Good Luck and don't be discouraged by the downvoters.<br>\nThey have nothing personal against you</p>",
      "votes": null,
      "replies": [
        {
          "id": 1573185,
          "author_name": "kalaikumarr",
          "author_url": "",
          "post_date": "11/06/2021 11:34:06",
          "content": "<p>I have corrected the heading after your comment. Thank you so much for your response and taking time to comment </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1573190,
      "author_name": "hamzaghanmi",
      "author_url": "",
      "post_date": "11/06/2021 11:39:27",
      "content": "<p>the official documentation of any library is enough, for example, if you want to use any functions built under pandas library you can check the documentation of that function on Pandas official website and you'll find an explication of all arguments + some code examples of using it (for ex the usage of <a href=\"https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.groupby.html\" target=\"_blank\">groupby</a>), and the same with any other library like seaborn, scikit learn ,numpy …</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1615430,
      "author_name": "carolineecc",
      "author_url": "",
      "post_date": "12/12/2021 08:34:09",
      "content": "<p>I learned python and machine learning from MOOC. In those courses, there are plenty of case studies, some are quite similar to Kaggle competiton's problem statement. We just need to learning by doing.;)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1572868": "I know Kaggle is a great plateform for ML, but I wonder how people are getting deep learning sample code for for their competitions. \n\nFor example I need sample code for library 1 but many time it is ending in error or due to outdated keyword. I know those baseline code is built using someother dataset, but if I alter the code for my dataset its getting error in most of the case. Can I know  \n\n1. How you are Getting the deep learning sample code ?\n1.1. In Kaggle we have plenty of code but is their any platform apart from kaggle like github repo, etc..\n\n2. Are you memorizing the library's keyword  and building the code from scratch?\n\n3. Are you modifying the existing code?. \n\n\nIf I am anything wrong please apologize me. Thanks in advance",
    "1573084": "Why do you need to memorize ? Normally people look up documentation when they write code. Most of the library comes with one.",
    "1573167": "____\n\n**I just want to point something important:** The reason people downvoted this post is not because of the question itself. \nThis is mainly because of the culture of Kaggle. [This is my own opinion that might be wrong] \n\nPeople downvote this post despite it containing a legitimate question simply because the title doesn't detail the question itself ['Help! Help! Help!']. I get it. you seek some help from the community and there are many different communities where the title 'Help! Help! Help!' will be accepted.  I think that editing it to something like: \"Help! Help! Help! Where do you get code examples from?\" will encourage people not only to upvote it but to even come and help. since it still has your \"personality in it 'Help! Help! Help!' but also it fits the kaggle culture of meaningful posts.   \n\n____\n\n**To the question:** By far the best method to get code examples for kaggle is.. well.. from kaggle. If you can't find what you look for in this competition (since it is new.. it takes time) you can always go back to previous related competitions. Often times you can find a list detailing past relevant competitions on the discussions. Just pay them a visit. You will find plenty of working end2end code on them. This is by far the best method for getting new ideas and pieces of code for kaggle. [In my opinion: Often for real-world machine learning projects as well] \n\n____\n\nGood Luck and don't be discouraged by the downvoters.\nThey have nothing personal against you",
    "1573185": "I have corrected the heading after your comment. Thank you so much for your response and taking time to comment",
    "1573190": "the official documentation of any library is enough, for example, if you want to use any functions built under pandas library you can check the documentation of that function on Pandas official website and you'll find an explication of all arguments + some code examples of using it (for ex the usage of [groupby](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.groupby.html)), and the same with any other library like seaborn, scikit learn ,numpy ...",
    "1615430": "I learned python and machine learning from MOOC. In those courses, there are plenty of case studies, some are quite similar to Kaggle competiton's problem statement. We just need to learning by doing.;)"
  },
  "source": "meta"
}