{
  "id": 21159,
  "title": "Place_id and location coordinates",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21159",
  "author_name": "",
  "post_date": "2016-05-23T20:25:28.017Z",
  "votes": null,
  "comment_count": 2,
  "views": 631,
  "content": "<p>Hi,</p>\n\n<p>I looked at the objective of what is to be predicted and the data (both train and test). So, the place_id needs to be predicted based on given x,y coordinates, timestamp and accuracy. I am not sure if I have understood this clearly. As per the train data, each place_id has a range set of x and y coordinates. Obviously,this range set of x,y coordinates are unique to each place_id. So, if the x and y coordinates are already provided in the test data, the place_id can be found out. So, what is the purpose of timestamp and accuracy? In other words, whatever the time and accuracy of the location coordinates, if I know the x and y coordinates, I can get the place_id. Its a simple mapping or am I missing  a big piece here?</p>\n\n<p>Thanks</p>",
  "messages": [
    {
      "id": "121098",
      "postDate": "05/23/2016 20:25:28",
      "content": "<p>Hi,</p>\n\n<p>I looked at the objective of what is to be predicted and the data (both train and test). So, the place_id needs to be predicted based on given x,y coordinates, timestamp and accuracy. I am not sure if I have understood this clearly. As per the train data, each place_id has a range set of x and y coordinates. Obviously,this range set of x,y coordinates are unique to each place_id. So, if the x and y coordinates are already provided in the test data, the place_id can be found out. So, what is the purpose of timestamp and accuracy? In other words, whatever the time and accuracy of the location coordinates, if I know the x and y coordinates, I can get the place_id. Its a simple mapping or am I missing  a big piece here?</p>\n\n<p>Thanks</p>",
      "rawMarkdown": "Hi,\r\n\r\nI looked at the objective of what is to be predicted and the data (both train and test). So, the place_id needs to be predicted based on given x,y coordinates, timestamp and accuracy. I am not sure if I have understood this clearly. As per the train data, each place_id has a range set of x and y coordinates. Obviously,this range set of x,y coordinates are unique to each place_id. So, if the x and y coordinates are already provided in the test data, the place_id can be found out. So, what is the purpose of timestamp and accuracy? In other words, whatever the time and accuracy of the location coordinates, if I know the x and y coordinates, I can get the place_id. Its a simple mapping or am I missing  a big piece here?\r\n\r\nThanks",
      "votes": null
    },
    {
      "id": "121099",
      "postDate": "05/23/2016 20:31:47",
      "content": "<p>Places overlap in their spatial distribution. Some of the scripts (my own included) show maps of the locations; in these maps, you can see that some of the phone locations are in very similar places, but belong to different place_id's.</p>\n\n<p>The key is that these are locations of the cell phones that are associated with a place. Two businesses might not overlap (or can they? think of multiple floors in a building ... we aren't given elevation!), but the coordinates of phones &quot;interested in&quot; those locations might.</p>",
      "rawMarkdown": "Places overlap in their spatial distribution. Some of the scripts (my own included) show maps of the locations; in these maps, you can see that some of the phone locations are in very similar places, but belong to different place_id's.\r\n\r\nThe key is that these are locations of the cell phones that are associated with a place. Two businesses might not overlap (or can they? think of multiple floors in a building ... we aren't given elevation!), but the coordinates of phones \"interested in\" those locations might.",
      "votes": null
    },
    {
      "id": "121100",
      "postDate": "05/23/2016 21:12:36",
      "content": "<p>That makes perfect sense. Thank you. </p>",
      "rawMarkdown": "That makes perfect sense. Thank you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 121099,
      "author_name": "zeraes",
      "author_url": "",
      "post_date": "05/23/2016 20:31:47",
      "content": "<p>Places overlap in their spatial distribution. Some of the scripts (my own included) show maps of the locations; in these maps, you can see that some of the phone locations are in very similar places, but belong to different place_id's.</p>\n\n<p>The key is that these are locations of the cell phones that are associated with a place. Two businesses might not overlap (or can they? think of multiple floors in a building ... we aren't given elevation!), but the coordinates of phones &quot;interested in&quot; those locations might.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 121100,
      "author_name": "timewarper",
      "author_url": "",
      "post_date": "05/23/2016 21:12:36",
      "content": "<p>That makes perfect sense. Thank you. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "121098": "Hi,\r\n\r\nI looked at the objective of what is to be predicted and the data (both train and test). So, the place_id needs to be predicted based on given x,y coordinates, timestamp and accuracy. I am not sure if I have understood this clearly. As per the train data, each place_id has a range set of x and y coordinates. Obviously,this range set of x,y coordinates are unique to each place_id. So, if the x and y coordinates are already provided in the test data, the place_id can be found out. So, what is the purpose of timestamp and accuracy? In other words, whatever the time and accuracy of the location coordinates, if I know the x and y coordinates, I can get the place_id. Its a simple mapping or am I missing  a big piece here?\r\n\r\nThanks",
    "121099": "Places overlap in their spatial distribution. Some of the scripts (my own included) show maps of the locations; in these maps, you can see that some of the phone locations are in very similar places, but belong to different place_id's.\r\n\r\nThe key is that these are locations of the cell phones that are associated with a place. Two businesses might not overlap (or can they? think of multiple floors in a building ... we aren't given elevation!), but the coordinates of phones \"interested in\" those locations might.",
    "121100": "That makes perfect sense. Thank you."
  },
  "source": "meta"
}