{
  "id": 20527,
  "title": "Comparing booking and non-booking hotel cluster distribution",
  "url": "/competitions/expedia-hotel-recommendations/discussion/20527",
  "author_name": "",
  "post_date": "2016-04-29T06:12:28.040Z",
  "votes": null,
  "comment_count": 2,
  "views": 617,
  "content": "<p>Here's a plot I made looking at the difference between the distribution of records within each hotel cluster for booking events and non-booking events in the training set. If people find this kind of thing useful, I'll try to make other plots that could be of interest in examining the training set.</p>\n\n<p>Legend:\nOrange: Booking Events\nBlue: Non-booking Events</p>",
  "messages": [
    {
      "id": "117452",
      "postDate": "04/29/2016 06:12:28",
      "content": "<p>Here's a plot I made looking at the difference between the distribution of records within each hotel cluster for booking events and non-booking events in the training set. If people find this kind of thing useful, I'll try to make other plots that could be of interest in examining the training set.</p>\n\n<p>Legend:\nOrange: Booking Events\nBlue: Non-booking Events</p>",
      "rawMarkdown": "Here's a plot I made looking at the difference between the distribution of records within each hotel cluster for booking events and non-booking events in the training set. If people find this kind of thing useful, I'll try to make other plots that could be of interest in examining the training set.\r\n\r\nLegend:\r\nOrange: Booking Events\r\nBlue: Non-booking Events",
      "votes": null
    },
    {
      "id": "117456",
      "postDate": "04/29/2016 06:44:28",
      "content": "<p>These information are always interesting. However, I think a bar chart might be more appropriate?</p>\n\n<p>Gerhard</p>",
      "rawMarkdown": "These information are always interesting. However, I think a bar chart might be more appropriate?\r\n\r\nGerhard",
      "votes": null
    },
    {
      "id": "117534",
      "postDate": "04/29/2016 14:51:01",
      "content": "<p>The issue with a bar chart is that in the program I'm using the percentages get stacked, so its harder to compare how similar the distributions are. If I figure out a way to unstack them I'll post a bar chart version!</p>",
      "rawMarkdown": "The issue with a bar chart is that in the program I'm using the percentages get stacked, so its harder to compare how similar the distributions are. If I figure out a way to unstack them I'll post a bar chart version!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 117456,
      "author_name": "mightybird",
      "author_url": "",
      "post_date": "04/29/2016 06:44:28",
      "content": "<p>These information are always interesting. However, I think a bar chart might be more appropriate?</p>\n\n<p>Gerhard</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 117534,
      "author_name": "nitini",
      "author_url": "",
      "post_date": "04/29/2016 14:51:01",
      "content": "<p>The issue with a bar chart is that in the program I'm using the percentages get stacked, so its harder to compare how similar the distributions are. If I figure out a way to unstack them I'll post a bar chart version!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "117452": "Here's a plot I made looking at the difference between the distribution of records within each hotel cluster for booking events and non-booking events in the training set. If people find this kind of thing useful, I'll try to make other plots that could be of interest in examining the training set.\r\n\r\nLegend:\r\nOrange: Booking Events\r\nBlue: Non-booking Events",
    "117456": "These information are always interesting. However, I think a bar chart might be more appropriate?\r\n\r\nGerhard",
    "117534": "The issue with a bar chart is that in the program I'm using the percentages get stacked, so its harder to compare how similar the distributions are. If I figure out a way to unstack them I'll post a bar chart version!"
  },
  "source": "meta"
}