{
  "id": 3232,
  "title": "flighthistory tailnumber vs icao_aircraft_type_actual",
  "url": "/competitions/flight/discussion/3232",
  "author_name": "",
  "post_date": "2012-12-01T21:51:29.033Z",
  "votes": null,
  "comment_count": 2,
  "views": 2840,
  "content": "<p>Which aircraft's tail number is in the field tailnumber? The originally scheduled aircraft or the actually flown one?<br>\r\nI'm asking because several flights have conflicting icao_aircraft_type_actual for the same tailnumber. One example is tail number N954U can be &quot;MD82&quot; or &quot;MD83&quot; in InitialTrainingSet/2012_11_12/FlightHistory/flighthistory.csv.</p>",
  "messages": [
    {
      "id": "17328",
      "postDate": "12/01/2012 21:51:29",
      "content": "<p>Which aircraft's tail number is in the field tailnumber? The originally scheduled aircraft or the actually flown one?<br>\r\nI'm asking because several flights have conflicting icao_aircraft_type_actual for the same tailnumber. One example is tail number N954U can be &quot;MD82&quot; or &quot;MD83&quot; in InitialTrainingSet/2012_11_12/FlightHistory/flighthistory.csv.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "17429",
      "postDate": "12/05/2012 19:39:22",
      "content": "<p>The field tailnumber in flighthistory.csv represents the final tailnumber seen for the flight in flighthistoryevents.csv, so it should generally correspond to the actual aircraft flown.</p>\r\n<p>Flighthistory.csv was a summary file FlightStats provided us along with the source (flighthistoryevents.csv) that we then downsampled and split according to day, so in general refer to flighthistoryevents.csv for any discrepancies that arise. Since the field\r\n tailnumber potentially leaks information, I will be removing it from flighthistory.csv for the test release.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "17454",
      "postDate": "12/06/2012 14:26:56",
      "content": "<p>[quote=Ben Hamner;17429]</p>\r\n<p>Since the field tailnumber potentially leaks information, I will be removing it from flighthistory.csv for the test release.</p>\r\n<p>[/quote]</p>\r\n<p>What information does it leak? True, it's easy to game the public leaderboard, but surely there are other ways to do that.</p>\r\n<p>Still, I think aircraft_id is is better for the purpose of identifying the plane, so tailnumber has no use.?</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 17429,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "12/05/2012 19:39:22",
      "content": "<p>The field tailnumber in flighthistory.csv represents the final tailnumber seen for the flight in flighthistoryevents.csv, so it should generally correspond to the actual aircraft flown.</p>\r\n<p>Flighthistory.csv was a summary file FlightStats provided us along with the source (flighthistoryevents.csv) that we then downsampled and split according to day, so in general refer to flighthistoryevents.csv for any discrepancies that arise. Since the field\r\n tailnumber potentially leaks information, I will be removing it from flighthistory.csv for the test release.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 17454,
      "author_name": "melisgl",
      "author_url": "",
      "post_date": "12/06/2012 14:26:56",
      "content": "<p>[quote=Ben Hamner;17429]</p>\r\n<p>Since the field tailnumber potentially leaks information, I will be removing it from flighthistory.csv for the test release.</p>\r\n<p>[/quote]</p>\r\n<p>What information does it leak? True, it's easy to game the public leaderboard, but surely there are other ways to do that.</p>\r\n<p>Still, I think aircraft_id is is better for the purpose of identifying the plane, so tailnumber has no use.?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "17328": "",
    "17429": "",
    "17454": ""
  },
  "source": "meta"
}