{
  "id": 278338,
  "title": "Contextualizing speed/acceleration of the football",
  "url": "/competitions/nfl-big-data-bowl-2022/discussion/278338",
  "author_name": "",
  "post_date": "2021-10-13T19:53:48.665444300Z",
  "votes": 3,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I noticed in the <code>tracking</code> datasets, there is no orientation for the direction of the football, yet there are speeds and accelerations. I'm wondering if these measurements are taken in the direction since the previous frame, or if there is a way to recover the directional component of these quantities. I'm interested in doing some physical analyses of the trajectories, but want to be sure I understand the context of the data better before doing so. Any help is appreciated!</p>",
  "messages": [
    {
      "id": "1543780",
      "postDate": "10/13/2021 19:53:48",
      "content": "<p>I noticed in the <code>tracking</code> datasets, there is no orientation for the direction of the football, yet there are speeds and accelerations. I'm wondering if these measurements are taken in the direction since the previous frame, or if there is a way to recover the directional component of these quantities. I'm interested in doing some physical analyses of the trajectories, but want to be sure I understand the context of the data better before doing so. Any help is appreciated!</p>",
      "rawMarkdown": "I noticed in the `tracking` datasets, there is no orientation for the direction of the football, yet there are speeds and accelerations. I'm wondering if these measurements are taken in the direction since the previous frame, or if there is a way to recover the directional component of these quantities. I'm interested in doing some physical analyses of the trajectories, but want to be sure I understand the context of the data better before doing so. Any help is appreciated!",
      "votes": null
    },
    {
      "id": "1543793",
      "postDate": "10/13/2021 20:23:58",
      "content": "<p>Great question. You can get the direction variable applying trigonometry to the change in x and y. Here is some pseudo code to help explain:</p>\n<p>xChange = x - lag(x)<br>\nyChange = y - lag(y)<br>\ndirection = arctan(yChange , xChange)</p>\n<p>Depending on which language /  packages you are using, you will likely need to add some factor of π to align the value with the direction used in the tracking data coordinates.</p>\n<p>Let me know if this helps!</p>",
      "rawMarkdown": "Great question. You can get the direction variable applying trigonometry to the change in x and y. Here is some pseudo code to help explain:\n\nxChange = x - lag(x)\nyChange = y - lag(y)\ndirection = arctan(yChange , xChange)\n\nDepending on which language /  packages you are using, you will likely need to add some factor of π to align the value with the direction used in the tracking data coordinates.\n\nLet me know if this helps!",
      "votes": null
    },
    {
      "id": "1572687",
      "postDate": "11/05/2021 21:25:21",
      "content": "<p><a href=\"https://www.kaggle.com/tombliss\" target=\"_blank\">@tombliss</a> A follow-up question on the speed of the football: </p>\n<p>Does the speed of the football in the tracking data include the vertical component? Or is this speed only in 2D (X/Y plane) similar to the players? </p>\n<p>Thanks!</p>",
      "rawMarkdown": "tombliss A follow-up question on the speed of the football: \n\nDoes the speed of the football in the tracking data include the vertical component? Or is this speed only in 2D (X/Y plane) similar to the players? \n\nThanks!",
      "votes": null
    },
    {
      "id": "1575439",
      "postDate": "11/08/2021 12:02:10",
      "content": "<p>It does not contain information about the vertical component. All of the tracking data only contains information in the horizontal x-y plane.</p>",
      "rawMarkdown": "It does not contain information about the vertical component. All of the tracking data only contains information in the horizontal x-y plane.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1543793,
      "author_name": "tombliss",
      "author_url": "",
      "post_date": "10/13/2021 20:23:58",
      "content": "<p>Great question. You can get the direction variable applying trigonometry to the change in x and y. Here is some pseudo code to help explain:</p>\n<p>xChange = x - lag(x)<br>\nyChange = y - lag(y)<br>\ndirection = arctan(yChange , xChange)</p>\n<p>Depending on which language /  packages you are using, you will likely need to add some factor of π to align the value with the direction used in the tracking data coordinates.</p>\n<p>Let me know if this helps!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1572687,
      "author_name": "jgeiman",
      "author_url": "",
      "post_date": "11/05/2021 21:25:21",
      "content": "<p><a href=\"https://www.kaggle.com/tombliss\" target=\"_blank\">@tombliss</a> A follow-up question on the speed of the football: </p>\n<p>Does the speed of the football in the tracking data include the vertical component? Or is this speed only in 2D (X/Y plane) similar to the players? </p>\n<p>Thanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1575439,
          "author_name": "tombliss",
          "author_url": "",
          "post_date": "11/08/2021 12:02:10",
          "content": "<p>It does not contain information about the vertical component. All of the tracking data only contains information in the horizontal x-y plane.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1543780": "I noticed in the `tracking` datasets, there is no orientation for the direction of the football, yet there are speeds and accelerations. I'm wondering if these measurements are taken in the direction since the previous frame, or if there is a way to recover the directional component of these quantities. I'm interested in doing some physical analyses of the trajectories, but want to be sure I understand the context of the data better before doing so. Any help is appreciated!",
    "1543793": "Great question. You can get the direction variable applying trigonometry to the change in x and y. Here is some pseudo code to help explain:\n\nxChange = x - lag(x)\nyChange = y - lag(y)\ndirection = arctan(yChange , xChange)\n\nDepending on which language /  packages you are using, you will likely need to add some factor of π to align the value with the direction used in the tracking data coordinates.\n\nLet me know if this helps!",
    "1572687": "tombliss A follow-up question on the speed of the football: \n\nDoes the speed of the football in the tracking data include the vertical component? Or is this speed only in 2D (X/Y plane) similar to the players? \n\nThanks!",
    "1575439": "It does not contain information about the vertical component. All of the tracking data only contains information in the horizontal x-y plane."
  },
  "source": "meta"
}