{
  "id": 542184,
  "title": "Assumptions on time_id",
  "url": "/competitions/jane-street-real-time-market-data-forecasting/discussion/542184",
  "author_name": "",
  "post_date": "2024-10-23T11:41:40.505754Z",
  "votes": 9,
  "comment_count": 2,
  "views": 0,
  "content": "<p>In the datascript description we are told</p>\n<blockquote>\n  <p>date_id and time_id - Integer values that are ordinally sorted, providing a chronological structure to the data, although the actual time intervals between time_id values may vary.</p>\n</blockquote>\n<p>Does this mean that the delta between time_id(0) and time_id(1) may not be the same as time_id(2) and time_id(3) or does it mean we can make no assumptions about the time_id's at all?</p>\n<p>For a hypothetical, could we consider time_id == 0 to be taken at the same time every day? Or is it just the first record for the day but on one day it might represent 9am and on another it might be 2pm or 4am?</p>\n<p>Essentially trying to understand if there may be signal in the values somewhere - e.g. one of the time_id's might be market open or market close</p>",
  "messages": [
    {
      "id": "3026032",
      "postDate": "10/23/2024 11:41:40",
      "content": "<p>In the datascript description we are told</p>\n<blockquote>\n  <p>date_id and time_id - Integer values that are ordinally sorted, providing a chronological structure to the data, although the actual time intervals between time_id values may vary.</p>\n</blockquote>\n<p>Does this mean that the delta between time_id(0) and time_id(1) may not be the same as time_id(2) and time_id(3) or does it mean we can make no assumptions about the time_id's at all?</p>\n<p>For a hypothetical, could we consider time_id == 0 to be taken at the same time every day? Or is it just the first record for the day but on one day it might represent 9am and on another it might be 2pm or 4am?</p>\n<p>Essentially trying to understand if there may be signal in the values somewhere - e.g. one of the time_id's might be market open or market close</p>",
      "rawMarkdown": "In the datascript description we are told\n\n> date_id and time_id - Integer values that are ordinally sorted, providing a chronological structure to the data, although the actual time intervals between time_id values may vary.\n\nDoes this mean that the delta between time_id(0) and time_id(1) may not be the same as time_id(2) and time_id(3) or does it mean we can make no assumptions about the time_id's at all?\n\nFor a hypothetical, could we consider time_id == 0 to be taken at the same time every day? Or is it just the first record for the day but on one day it might represent 9am and on another it might be 2pm or 4am?\n\nEssentially trying to understand if there may be signal in the values somewhere - e.g. one of the time_id's might be market open or market close",
      "votes": null
    },
    {
      "id": "3027092",
      "postDate": "10/24/2024 14:10:54",
      "content": "<p>It would be great if the hosts at leasts could align the discrepancies. Some days have more time_ids than others.</p>\n<p>I don't think it matters if we know what time_id = x means, but it would definitely help if we know how to map one time_id to the other.</p>",
      "rawMarkdown": "It would be great if the hosts at leasts could align the discrepancies. Some days have more time_ids than others.\n\nI don't think it matters if we know what time_id = x means, but it would definitely help if we know how to map one time_id to the other.",
      "votes": null
    },
    {
      "id": "3037490",
      "postDate": "11/05/2024 18:42:45",
      "content": "<p>The host has partially replied here: <a href=\"https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206\" target=\"_blank\">https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206</a>.</p>",
      "rawMarkdown": "The host has partially replied here: https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3027092,
      "author_name": "verracodeguacas",
      "author_url": "",
      "post_date": "10/24/2024 14:10:54",
      "content": "<p>It would be great if the hosts at leasts could align the discrepancies. Some days have more time_ids than others.</p>\n<p>I don't think it matters if we know what time_id = x means, but it would definitely help if we know how to map one time_id to the other.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3037490,
      "author_name": "aleixgim",
      "author_url": "",
      "post_date": "11/05/2024 18:42:45",
      "content": "<p>The host has partially replied here: <a href=\"https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206\" target=\"_blank\">https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206</a>.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3026032": "In the datascript description we are told\n\n> date_id and time_id - Integer values that are ordinally sorted, providing a chronological structure to the data, although the actual time intervals between time_id values may vary.\n\nDoes this mean that the delta between time_id(0) and time_id(1) may not be the same as time_id(2) and time_id(3) or does it mean we can make no assumptions about the time_id's at all?\n\nFor a hypothetical, could we consider time_id == 0 to be taken at the same time every day? Or is it just the first record for the day but on one day it might represent 9am and on another it might be 2pm or 4am?\n\nEssentially trying to understand if there may be signal in the values somewhere - e.g. one of the time_id's might be market open or market close",
    "3027092": "It would be great if the hosts at leasts could align the discrepancies. Some days have more time_ids than others.\n\nI don't think it matters if we know what time_id = x means, but it would definitely help if we know how to map one time_id to the other.",
    "3037490": "The host has partially replied here: https://www.kaggle.com/competitions/jane-street-real-time-market-data-forecasting/discussion/540436#3036206."
  },
  "source": "meta"
}