{
  "id": 2695,
  "title": "full data set?",
  "url": "/competitions/acm-sf-chapter-hackathon-big/discussion/2695",
  "author_name": "",
  "post_date": "2012-09-18T18:38:25.877Z",
  "votes": null,
  "comment_count": 3,
  "views": 4017,
  "content": "<p>The posted data 'large' set seems to only be a small portion of the described full data set? &nbsp;It looks like it only has data from August through October of 2011, with about 1.8 million clicks and 1.2 million users? &nbsp;The task description talks about 2 years\r\n of data with 67 million clicks and 8 million users. &nbsp;</p>\r\n<p>The product data is almost 8GB, so the overall size meets the description - but there seem to be much fewer events. &nbsp;Did I misunderstand something about the description?</p>\r\n<p>Thanks for any clarification!</p>",
  "messages": [
    {
      "id": "14488",
      "postDate": "09/18/2012 18:38:25",
      "content": "<p>The posted data 'large' set seems to only be a small portion of the described full data set? &nbsp;It looks like it only has data from August through October of 2011, with about 1.8 million clicks and 1.2 million users? &nbsp;The task description talks about 2 years\r\n of data with 67 million clicks and 8 million users. &nbsp;</p>\r\n<p>The product data is almost 8GB, so the overall size meets the description - but there seem to be much fewer events. &nbsp;Did I misunderstand something about the description?</p>\r\n<p>Thanks for any clarification!</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14494",
      "postDate": "09/18/2012 19:17:03",
      "content": "<p>It is indeed the full dataset. &nbsp;The description was written by someone else before the start of the competition while I was still trying to wriggle the dataset out of the business. &nbsp;The numbers were from the top of my head on a very early version of the\r\n dataset that wasn't scrubbed. &nbsp; Was just a timing problem on our end. :-)</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14498",
      "postDate": "09/18/2012 20:08:24",
      "content": "<p>Thanks, that helps make sense. &nbsp;Might be good to update the description?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14499",
      "postDate": "09/18/2012 20:08:50",
      "content": "<p>a user clicks only one product on a search results page .</p>\r\n<p>Why would you want us to predict 5 products for a query? <br>\r\nGiven that the history exists for only one product, what information are you using to get MAP metric?</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 14494,
      "author_name": "nickkolegraff",
      "author_url": "",
      "post_date": "09/18/2012 19:17:03",
      "content": "<p>It is indeed the full dataset. &nbsp;The description was written by someone else before the start of the competition while I was still trying to wriggle the dataset out of the business. &nbsp;The numbers were from the top of my head on a very early version of the\r\n dataset that wasn't scrubbed. &nbsp; Was just a timing problem on our end. :-)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14498,
      "author_name": "irreverent",
      "author_url": "",
      "post_date": "09/18/2012 20:08:24",
      "content": "<p>Thanks, that helps make sense. &nbsp;Might be good to update the description?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14499,
      "author_name": "rkirana",
      "author_url": "",
      "post_date": "09/18/2012 20:08:50",
      "content": "<p>a user clicks only one product on a search results page .</p>\r\n<p>Why would you want us to predict 5 products for a query? <br>\r\nGiven that the history exists for only one product, what information are you using to get MAP metric?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "14488": "",
    "14494": "",
    "14498": "",
    "14499": ""
  },
  "source": "meta"
}