{
  "id": 20883,
  "title": "Place location != Check-in Location",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/20883",
  "author_name": "🔥🦫📿🗽",
  "post_date": "2016-05-12T04:39:07.887000",
  "votes": 0,
  "comment_count": 0,
  "views": 851,
  "content": "<p>Poor me, I have no Facebook account and don't do check-ins...</p>\n\n<p>So, I made the assumption that the place location is close to the check-in location and I basically took the average of all (X, Y) for a given place ID as the place location.</p>\n\n<p>Calculating the min, max, avg, std distance from so calculated place location from the training data suggests that this is NOT a very helpful thing to do. Of course, this varies from place to place.</p>\n\n<p>So, it seems that people are checking-in from locations from home, work or other locations, rather than from close or inside the place ID or the accuracy of user location is quite variable.</p>",
  "messages": [
    {
      "id": 119659,
      "postDate": "2016-05-12T04:39:07.887Z",
      "content": "<p>Poor me, I have no Facebook account and don't do check-ins...</p>\n\n<p>So, I made the assumption that the place location is close to the check-in location and I basically took the average of all (X, Y) for a given place ID as the place location.</p>\n\n<p>Calculating the min, max, avg, std distance from so calculated place location from the training data suggests that this is NOT a very helpful thing to do. Of course, this varies from place to place.</p>\n\n<p>So, it seems that people are checking-in from locations from home, work or other locations, rather than from close or inside the place ID or the accuracy of user location is quite variable.</p>",
      "rawMarkdown": "Poor me, I have no Facebook account and don't do check-ins...\r\n\r\nSo, I made the assumption that the place location is close to the check-in location and I basically took the average of all (X, Y) for a given place ID as the place location.\r\n\r\nCalculating the min, max, avg, std distance from so calculated place location from the training data suggests that this is NOT a very helpful thing to do. Of course, this varies from place to place.\r\n\r\nSo, it seems that people are checking-in from locations from home, work or other locations, rather than from close or inside the place ID or the accuracy of user location is quite variable."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "119659": "Poor me, I have no Facebook account and don't do check-ins...\r\n\r\nSo, I made the assumption that the place location is close to the check-in location and I basically took the average of all (X, Y) for a given place ID as the place location.\r\n\r\nCalculating the min, max, avg, std distance from so calculated place location from the training data suggests that this is NOT a very helpful thing to do. Of course, this varies from place to place.\r\n\r\nSo, it seems that people are checking-in from locations from home, work or other locations, rather than from close or inside the place ID or the accuracy of user location is quite variable."
  }
}