{
  "id": 93503,
  "title": "16 Earthquakes visualized from a different point of view",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/93503",
  "author_name": "",
  "post_date": "2019-05-27T17:14:47.070948400Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Most of the people I have found are predicting the time following a periodic function (zig zag/ sawtooth in time to next earthquake plots). </p>\n\n<p>Introduction to periodic and aperiodic signals: <img src=\"http://3.bp.blogspot.com/-TCfkXTROiRo/VBxrHylX0xI/AAAAAAAAANc/X45cASgFUt0/s1600/periodic.png\" alt=\"\"></p>\n\n<p>I am finding it a bit tricky to believe that earthquakes are periodic. It would mean they represent a deterministic process. For example an earthquake occurs after every x unit of time at an epicenter (for sure).</p>\n\n<p>After wrangling the dataset I found a pattern which looks more <strong>realistic</strong> and captures the seismic activity of an actual earthquake and aftershocks. One another interesting pattern which it shows is shockwaves are clustered here representing a stronger seismic activity unlike the periodic pattern which appears to be discrete (hence non realistic). Also if we see the publications by authors they have accounted for kurtosis and variance which can be seen in this point of view of earthquakes.</p>\n\n<p>I haven't seen this plot in the competition or any discussion regarding it so far so I thought to share. </p>\n\n<p><strong>Note: This is not a sample. It is entire dataset shown in the visualization.</strong>\n<img src=\"http://i63.tinypic.com/o7pouh.png\" alt=\"\"></p>",
  "messages": [
    {
      "id": "537826",
      "postDate": "05/27/2019 17:14:47",
      "content": "<p>Most of the people I have found are predicting the time following a periodic function (zig zag/ sawtooth in time to next earthquake plots). </p>\n\n<p>Introduction to periodic and aperiodic signals: <img src=\"http://3.bp.blogspot.com/-TCfkXTROiRo/VBxrHylX0xI/AAAAAAAAANc/X45cASgFUt0/s1600/periodic.png\" alt=\"\"></p>\n\n<p>I am finding it a bit tricky to believe that earthquakes are periodic. It would mean they represent a deterministic process. For example an earthquake occurs after every x unit of time at an epicenter (for sure).</p>\n\n<p>After wrangling the dataset I found a pattern which looks more <strong>realistic</strong> and captures the seismic activity of an actual earthquake and aftershocks. One another interesting pattern which it shows is shockwaves are clustered here representing a stronger seismic activity unlike the periodic pattern which appears to be discrete (hence non realistic). Also if we see the publications by authors they have accounted for kurtosis and variance which can be seen in this point of view of earthquakes.</p>\n\n<p>I haven't seen this plot in the competition or any discussion regarding it so far so I thought to share. </p>\n\n<p><strong>Note: This is not a sample. It is entire dataset shown in the visualization.</strong>\n<img src=\"http://i63.tinypic.com/o7pouh.png\" alt=\"\"></p>",
      "rawMarkdown": "Most of the people I have found are predicting the time following a periodic function (zig zag/ sawtooth in time to next earthquake plots). \n\nIntroduction to periodic and aperiodic signals: ![](http://3.bp.blogspot.com/-TCfkXTROiRo/VBxrHylX0xI/AAAAAAAAANc/X45cASgFUt0/s1600/periodic.png)\n\nI am finding it a bit tricky to believe that earthquakes are periodic. It would mean they represent a deterministic process. For example an earthquake occurs after every x unit of time at an epicenter (for sure).\n\nAfter wrangling the dataset I found a pattern which looks more **realistic** and captures the seismic activity of an actual earthquake and aftershocks. One another interesting pattern which it shows is shockwaves are clustered here representing a stronger seismic activity unlike the periodic pattern which appears to be discrete (hence non realistic). Also if we see the publications by authors they have accounted for kurtosis and variance which can be seen in this point of view of earthquakes.\n\nI haven't seen this plot in the competition or any discussion regarding it so far so I thought to share. \n\n**Note: This is not a sample. It is entire dataset shown in the visualization.**\n![](http://i63.tinypic.com/o7pouh.png)",
      "votes": null
    },
    {
      "id": "538347",
      "postDate": "05/28/2019 12:53:32",
      "content": "<p>Can you show this chart for both test and train.</p>",
      "rawMarkdown": "Can you show this chart for both test and train.",
      "votes": null
    },
    {
      "id": "538350",
      "postDate": "05/28/2019 12:56:14",
      "content": "<p>Peak detection is therefore extremely important</p>",
      "rawMarkdown": "Peak detection is therefore extremely important",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 538347,
      "author_name": "hmcranbercourt",
      "author_url": "",
      "post_date": "05/28/2019 12:53:32",
      "content": "<p>Can you show this chart for both test and train.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 538350,
      "author_name": "hmcranbercourt",
      "author_url": "",
      "post_date": "05/28/2019 12:56:14",
      "content": "<p>Peak detection is therefore extremely important</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "537826": "Most of the people I have found are predicting the time following a periodic function (zig zag/ sawtooth in time to next earthquake plots). \n\nIntroduction to periodic and aperiodic signals: ![](http://3.bp.blogspot.com/-TCfkXTROiRo/VBxrHylX0xI/AAAAAAAAANc/X45cASgFUt0/s1600/periodic.png)\n\nI am finding it a bit tricky to believe that earthquakes are periodic. It would mean they represent a deterministic process. For example an earthquake occurs after every x unit of time at an epicenter (for sure).\n\nAfter wrangling the dataset I found a pattern which looks more **realistic** and captures the seismic activity of an actual earthquake and aftershocks. One another interesting pattern which it shows is shockwaves are clustered here representing a stronger seismic activity unlike the periodic pattern which appears to be discrete (hence non realistic). Also if we see the publications by authors they have accounted for kurtosis and variance which can be seen in this point of view of earthquakes.\n\nI haven't seen this plot in the competition or any discussion regarding it so far so I thought to share. \n\n**Note: This is not a sample. It is entire dataset shown in the visualization.**\n![](http://i63.tinypic.com/o7pouh.png)",
    "538347": "Can you show this chart for both test and train.",
    "538350": "Peak detection is therefore extremely important"
  },
  "source": "meta"
}