{
  "id": 21328,
  "title": "2-D check-in scatter plot of several place_id",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21328",
  "author_name": "",
  "post_date": "2016-05-31T04:09:56.507Z",
  "votes": 2,
  "comment_count": 2,
  "views": 651,
  "content": "<p>I plot the 2-D check-in point distributions for several place_id. \nThe figures are attached in the post. The figure files are named as &quot;A-B.png&quot;. A means the number of check-in points, and B means the place_id.</p>\n\n<p>The x, y axises are the same as the training data's.\nThe one circle is one check-in record. The radius of the each circle is the accuracy value of that check-in record. From the examples I attach, we could observe several patterns.</p>\n\n<ol>\n<li><p>Two dense check-in centers.  1099-1654576295.png, 100-1013434352.png show two dense check-in centers.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/1099-1654576295.png\" alt=\"1099-1654576295\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-1013434352.png\" alt=\"100-1013434352\" title></p></li>\n<li><p>One dense center.  78-3564561498, 100-2090357921 only have one dense center. \n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/78-3564561498.png\" alt=\"78-3564561498\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-2090357921.png\" alt=\"100-2090357921\" title></p></li>\n</ol>\n\n<p>In both pattern 1, 2 it seems that the check in distribution is horizontal.</p>\n\n<ol start=\"3\">\n<li><p>2-D Gaussian distribution. 500-1037530880, 1099-2972810778 look like 2-D Gaussian distribution with very dense check-in points at the center.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/500-1037530880.png\" alt=\"500-1037530880\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/1099-2972810778.png\" alt=\"1099-2972810778\" title></p></li>\n<li><p>100-1008823061 shows random check-in patterns.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-1008823061.png\" alt=\"100-1008823061\" title></p></li>\n</ol>\n\n<p>I did not explore how &quot;time&quot; affect the check-in patterns. This is my next step. I hope these figures could help you on explore the data set.</p>",
  "messages": [
    {
      "id": "121933",
      "postDate": "05/31/2016 04:09:56",
      "content": "<p>I plot the 2-D check-in point distributions for several place_id. \nThe figures are attached in the post. The figure files are named as &quot;A-B.png&quot;. A means the number of check-in points, and B means the place_id.</p>\n\n<p>The x, y axises are the same as the training data's.\nThe one circle is one check-in record. The radius of the each circle is the accuracy value of that check-in record. From the examples I attach, we could observe several patterns.</p>\n\n<ol>\n<li><p>Two dense check-in centers.  1099-1654576295.png, 100-1013434352.png show two dense check-in centers.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/1099-1654576295.png\" alt=\"1099-1654576295\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-1013434352.png\" alt=\"100-1013434352\" title></p></li>\n<li><p>One dense center.  78-3564561498, 100-2090357921 only have one dense center. \n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/78-3564561498.png\" alt=\"78-3564561498\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-2090357921.png\" alt=\"100-2090357921\" title></p></li>\n</ol>\n\n<p>In both pattern 1, 2 it seems that the check in distribution is horizontal.</p>\n\n<ol start=\"3\">\n<li><p>2-D Gaussian distribution. 500-1037530880, 1099-2972810778 look like 2-D Gaussian distribution with very dense check-in points at the center.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/500-1037530880.png\" alt=\"500-1037530880\" title>\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/1099-2972810778.png\" alt=\"1099-2972810778\" title></p></li>\n<li><p>100-1008823061 shows random check-in patterns.\n<img src=\"http://www.cs.nmsu.edu/~chu/kaggle/100-1008823061.png\" alt=\"100-1008823061\" title></p></li>\n</ol>\n\n<p>I did not explore how &quot;time&quot; affect the check-in patterns. This is my next step. I hope these figures could help you on explore the data set.</p>",
      "rawMarkdown": "I plot the 2-D check-in point distributions for several place_id. \r\nThe figures are attached in the post. The figure files are named as \"A-B.png\". A means the number of check-in points, and B means the place_id.\r\n\r\nThe x, y axises are the same as the training data's.\r\nThe one circle is one check-in record. The radius of the each circle is the accuracy value of that check-in record. From the examples I attach, we could observe several patterns.\r\n\r\n 1. Two dense check-in centers.  1099-1654576295.png, 100-1013434352.png show two dense check-in centers.\r\n![1099-1654576295][1]\r\n![100-1013434352][2]\r\n\r\n 2. One dense center.  78-3564561498, 100-2090357921 only have one dense center. \r\n![78-3564561498][3]\r\n![100-2090357921][4]\r\n\r\nIn both pattern 1, 2 it seems that the check in distribution is horizontal.\r\n\r\n 3. 2-D Gaussian distribution. 500-1037530880, 1099-2972810778 look like 2-D Gaussian distribution with very dense check-in points at the center.\r\n![500-1037530880][5]\r\n![1099-2972810778][6]\r\n\r\n 4. 100-1008823061 shows random check-in patterns.\r\n![100-1008823061][7]\r\n\r\nI did not explore how \"time\" affect the check-in patterns. This is my next step. I hope these figures could help you on explore the data set.\r\n\r\n\r\n  [1]: http://www.cs.nmsu.edu/~chu/kaggle/1099-1654576295.png\r\n  [2]: http://www.cs.nmsu.edu/~chu/kaggle/100-1013434352.png\r\n  [3]: http://www.cs.nmsu.edu/~chu/kaggle/78-3564561498.png\r\n  [4]: http://www.cs.nmsu.edu/~chu/kaggle/100-2090357921.png\r\n  [5]: http://www.cs.nmsu.edu/~chu/kaggle/500-1037530880.png\r\n  [6]: http://www.cs.nmsu.edu/~chu/kaggle/1099-2972810778.png\r\n  [7]: http://www.cs.nmsu.edu/~chu/kaggle/100-1008823061.png",
      "votes": null
    },
    {
      "id": "121934",
      "postDate": "05/31/2016 04:26:38",
      "content": "<p>Hi, quick question. The x-axis is the <strong>x</strong> position of the place id or the <strong>y</strong> position?</p>",
      "rawMarkdown": "Hi, quick question. The x-axis is the **x** position of the place id or the **y** position?",
      "votes": null
    },
    {
      "id": "121938",
      "postDate": "05/31/2016 04:35:25",
      "content": "<p>Hi, x-axis is the 'x' column of the given data.</p>",
      "rawMarkdown": "Hi, x-axis is the 'x' column of the given data.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 121934,
      "author_name": "ma350365879",
      "author_url": "",
      "post_date": "05/31/2016 04:26:38",
      "content": "<p>Hi, quick question. The x-axis is the <strong>x</strong> position of the place id or the <strong>y</strong> position?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 121938,
      "author_name": "chu1990",
      "author_url": "",
      "post_date": "05/31/2016 04:35:25",
      "content": "<p>Hi, x-axis is the 'x' column of the given data.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "121933": "I plot the 2-D check-in point distributions for several place_id. \r\nThe figures are attached in the post. The figure files are named as \"A-B.png\". A means the number of check-in points, and B means the place_id.\r\n\r\nThe x, y axises are the same as the training data's.\r\nThe one circle is one check-in record. The radius of the each circle is the accuracy value of that check-in record. From the examples I attach, we could observe several patterns.\r\n\r\n 1. Two dense check-in centers.  1099-1654576295.png, 100-1013434352.png show two dense check-in centers.\r\n![1099-1654576295][1]\r\n![100-1013434352][2]\r\n\r\n 2. One dense center.  78-3564561498, 100-2090357921 only have one dense center. \r\n![78-3564561498][3]\r\n![100-2090357921][4]\r\n\r\nIn both pattern 1, 2 it seems that the check in distribution is horizontal.\r\n\r\n 3. 2-D Gaussian distribution. 500-1037530880, 1099-2972810778 look like 2-D Gaussian distribution with very dense check-in points at the center.\r\n![500-1037530880][5]\r\n![1099-2972810778][6]\r\n\r\n 4. 100-1008823061 shows random check-in patterns.\r\n![100-1008823061][7]\r\n\r\nI did not explore how \"time\" affect the check-in patterns. This is my next step. I hope these figures could help you on explore the data set.\r\n\r\n\r\n  [1]: http://www.cs.nmsu.edu/~chu/kaggle/1099-1654576295.png\r\n  [2]: http://www.cs.nmsu.edu/~chu/kaggle/100-1013434352.png\r\n  [3]: http://www.cs.nmsu.edu/~chu/kaggle/78-3564561498.png\r\n  [4]: http://www.cs.nmsu.edu/~chu/kaggle/100-2090357921.png\r\n  [5]: http://www.cs.nmsu.edu/~chu/kaggle/500-1037530880.png\r\n  [6]: http://www.cs.nmsu.edu/~chu/kaggle/1099-2972810778.png\r\n  [7]: http://www.cs.nmsu.edu/~chu/kaggle/100-1008823061.png",
    "121934": "Hi, quick question. The x-axis is the **x** position of the place id or the **y** position?",
    "121938": "Hi, x-axis is the 'x' column of the given data."
  },
  "source": "meta"
}