{
  "id": 20516,
  "title": "Question about impressions",
  "url": "/competitions/expedia-personalized-sort/discussion/20516",
  "author_name": "",
  "post_date": "2016-04-28T18:06:25.003Z",
  "votes": null,
  "comment_count": 2,
  "views": 730,
  "content": "<p>Hello, </p>\n\n<p>I am writing because I have a question about the impressions of the data. </p>\n\n<p>I wonder, in the train (or test) data sets, the impressions that each 'srch_id' viewed \nare somewhat customized in some way or completely randomized. \nIn the data set, we have 'srch_id', so I guess that the impressions that the \ncustomers viewed are somewhat customized based on their impression, click, purchase history or other individual sorting criteria. That is the hotel properties that the customers viewed reflect the customers' preferences or tastes in some sense. </p>\n\n<p>I found this interesting fact that the price dispersion of the hotels that each srch_id viewed significantly affects customers' booking decision. I assumed that the impressions that the customers viewed somewhat reflect the customers' taste, historical purchases or preferences in some sense and want to make sure if this is correct.</p>\n\n<p>Thank you for your help in advance and look forward to hearing back from you.</p>\n\n<p>Best regards,\nEunhee Emily Ko</p>",
  "messages": [
    {
      "id": "117360",
      "postDate": "04/28/2016 18:06:25",
      "content": "<p>Hello, </p>\n\n<p>I am writing because I have a question about the impressions of the data. </p>\n\n<p>I wonder, in the train (or test) data sets, the impressions that each 'srch_id' viewed \nare somewhat customized in some way or completely randomized. \nIn the data set, we have 'srch_id', so I guess that the impressions that the \ncustomers viewed are somewhat customized based on their impression, click, purchase history or other individual sorting criteria. That is the hotel properties that the customers viewed reflect the customers' preferences or tastes in some sense. </p>\n\n<p>I found this interesting fact that the price dispersion of the hotels that each srch_id viewed significantly affects customers' booking decision. I assumed that the impressions that the customers viewed somewhat reflect the customers' taste, historical purchases or preferences in some sense and want to make sure if this is correct.</p>\n\n<p>Thank you for your help in advance and look forward to hearing back from you.</p>\n\n<p>Best regards,\nEunhee Emily Ko</p>",
      "rawMarkdown": "Hello, \r\n\r\nI am writing because I have a question about the impressions of the data. \r\n\r\nI wonder, in the train (or test) data sets, the impressions that each 'srch_id' viewed \r\nare somewhat customized in some way or completely randomized. \r\nIn the data set, we have 'srch_id', so I guess that the impressions that the \r\ncustomers viewed are somewhat customized based on their impression, click, purchase history or other individual sorting criteria. That is the hotel properties that the customers viewed reflect the customers' preferences or tastes in some sense. \r\n\r\nI found this interesting fact that the price dispersion of the hotels that each srch_id viewed significantly affects customers' booking decision. I assumed that the impressions that the customers viewed somewhat reflect the customers' taste, historical purchases or preferences in some sense and want to make sure if this is correct.\r\n\r\nThank you for your help in advance and look forward to hearing back from you.\r\n\r\nBest regards,\r\nEunhee Emily Ko",
      "votes": null
    },
    {
      "id": "117499",
      "postDate": "04/29/2016 12:44:35",
      "content": "<p>If random_bool = 1 then an impression was randomized. Otherwise, an impression was generated by our internal algorithm.</p>\n\n<p>Adam</p>",
      "rawMarkdown": "If random_bool = 1 then an impression was randomized. Otherwise, an impression was generated by our internal algorithm.\r\n\r\nAdam",
      "votes": null
    },
    {
      "id": "125265",
      "postDate": "06/28/2016 08:50:29",
      "content": "<p>Hi Emily, </p>\n\n<p>I also want to use the data set for research purpose. Could you find exactly variable names; for example, &quot;prop_country_id=9&quot; is USA etc.</p>\n\n<p>Many thank.</p>\n\n<p>David</p>",
      "rawMarkdown": "Hi Emily, \r\n\r\nI also want to use the data set for research purpose. Could you find exactly variable names; for example, \"prop_country_id=9\" is USA etc.\r\n\r\nMany thank.\r\n\r\nDavid",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 117499,
      "author_name": "adamwoz",
      "author_url": "",
      "post_date": "04/29/2016 12:44:35",
      "content": "<p>If random_bool = 1 then an impression was randomized. Otherwise, an impression was generated by our internal algorithm.</p>\n\n<p>Adam</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 125265,
      "author_name": "davidtang",
      "author_url": "",
      "post_date": "06/28/2016 08:50:29",
      "content": "<p>Hi Emily, </p>\n\n<p>I also want to use the data set for research purpose. Could you find exactly variable names; for example, &quot;prop_country_id=9&quot; is USA etc.</p>\n\n<p>Many thank.</p>\n\n<p>David</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "117360": "Hello, \r\n\r\nI am writing because I have a question about the impressions of the data. \r\n\r\nI wonder, in the train (or test) data sets, the impressions that each 'srch_id' viewed \r\nare somewhat customized in some way or completely randomized. \r\nIn the data set, we have 'srch_id', so I guess that the impressions that the \r\ncustomers viewed are somewhat customized based on their impression, click, purchase history or other individual sorting criteria. That is the hotel properties that the customers viewed reflect the customers' preferences or tastes in some sense. \r\n\r\nI found this interesting fact that the price dispersion of the hotels that each srch_id viewed significantly affects customers' booking decision. I assumed that the impressions that the customers viewed somewhat reflect the customers' taste, historical purchases or preferences in some sense and want to make sure if this is correct.\r\n\r\nThank you for your help in advance and look forward to hearing back from you.\r\n\r\nBest regards,\r\nEunhee Emily Ko",
    "117499": "If random_bool = 1 then an impression was randomized. Otherwise, an impression was generated by our internal algorithm.\r\n\r\nAdam",
    "125265": "Hi Emily, \r\n\r\nI also want to use the data set for research purpose. Could you find exactly variable names; for example, \"prop_country_id=9\" is USA etc.\r\n\r\nMany thank.\r\n\r\nDavid"
  },
  "source": "meta"
}