{
  "id": 3520,
  "title": "Redirected/Diverted flights",
  "url": "/competitions/flight/discussion/3520",
  "author_name": "",
  "post_date": "2013-01-02T13:17:05.350Z",
  "votes": 5,
  "comment_count": 16,
  "views": 15080,
  "content": "<p>What are we supposed to do with redirected/diverted&nbsp;flights?</p>\r\n<p>The data for them seems to be a mess.</p>\r\n<p>For example, check out flight_history_id=281435994 on 11/26.</p>\r\n<p>From what I can tell, it was going from LGA to MKE, but was diverted to ROC. It stayed there for about an hour, and then went to MKE.</p>\r\n<p>The data in flighthistory.csv is screwy; the actual_gate_arrival is for MKE, but the actual_runway_arrival is for ROC (and off by an hour due to a timezone mixup, I think).</p>\r\n<p>I couldn't quickly find an example, but I assume it is possible for the flight to be diverted and never continue to the original destination airport.</p>\r\n<p>Given all this, a few specific questions:</p>\r\n<p>1) Are you intending to keep diverted flights in the test set?</p>\r\n<p>2) If so, are we supposed to be predicting the arrival times for the airport it was diverted to, or the original destination airport?</p>\r\n<p>3) Have you considered capping the error for a single flight when computing the score? From a practical standpoint, there isn't a ton of difference between being off by 2 hours or 3 hours; both are terrible, but the rmse will be very different.</p>\r\n<p>&nbsp;-Dave</p>",
  "messages": [
    {
      "id": "18892",
      "postDate": "01/02/2013 13:17:05",
      "content": "<p>What are we supposed to do with redirected/diverted&nbsp;flights?</p>\r\n<p>The data for them seems to be a mess.</p>\r\n<p>For example, check out flight_history_id=281435994 on 11/26.</p>\r\n<p>From what I can tell, it was going from LGA to MKE, but was diverted to ROC. It stayed there for about an hour, and then went to MKE.</p>\r\n<p>The data in flighthistory.csv is screwy; the actual_gate_arrival is for MKE, but the actual_runway_arrival is for ROC (and off by an hour due to a timezone mixup, I think).</p>\r\n<p>I couldn't quickly find an example, but I assume it is possible for the flight to be diverted and never continue to the original destination airport.</p>\r\n<p>Given all this, a few specific questions:</p>\r\n<p>1) Are you intending to keep diverted flights in the test set?</p>\r\n<p>2) If so, are we supposed to be predicting the arrival times for the airport it was diverted to, or the original destination airport?</p>\r\n<p>3) Have you considered capping the error for a single flight when computing the score? From a practical standpoint, there isn't a ton of difference between being off by 2 hours or 3 hours; both are terrible, but the rmse will be very different.</p>\r\n<p>&nbsp;-Dave</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "18902",
      "postDate": "01/02/2013 17:04:44",
      "content": "<p>[quote=Dave;18892]</p>\r\n<p>What are we supposed to do with redirected/diverted&nbsp;flights?</p>\r\n<p>The data for them seems to be a mess.</p>\r\n<p>For example, check out flight_history_id=281435994 on 11/26.</p>\r\n<p>From what I can tell, it was going from LGA to MKE, but was diverted to ROC. It stayed there for about an hour, and then went to MKE.</p>\r\n<p>The data in flighthistory.csv is screwy; the actual_gate_arrival is for MKE, but the actual_runway_arrival is for ROC (and off by an hour due to a timezone mixup, I think).</p>\r\n<p>I couldn't quickly find an example, but I assume it is possible for the flight to be diverted and never continue to the original destination airport.</p>\r\n<p>Given all this, a few specific questions:</p>\r\n<p>1) Are you intending to keep diverted flights in the test set?</p>\r\n<p>2) If so, are we supposed to be predicting the arrival times for the airport it was diverted to, or the original destination airport?</p>\r\n<p>3) Have you considered capping the error for a single flight when computing the score? From a practical standpoint, there isn't a ton of difference between being off by 2 hours or 3 hours; both are terrible, but the rmse will be very different.</p>\r\n<p>&nbsp;-Dave</p>\r\n<p>[/quote]There shouldn't be any diverted flights in the public leaderboard set that you're submitting predictions on, and you won't be asked to predict any diverted flights in the final evaluation set.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "18907",
      "postDate": "01/02/2013 17:32:01",
      "content": "<p>I think flight_history_id=281453856 on 11/26 is in the test set and was diverted.</p>\r\n<p>281453856,2012-11-26 05:19:39.373000-08:00,STATUS-Diverted,&quot;AGD- New=11/26/12 07:55, STATUS- Old=R New=D&quot;</p>\r\n<p>&nbsp; Dave</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19094",
      "postDate": "01/08/2013 15:22:11",
      "content": "<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19107",
      "postDate": "01/08/2013 19:27:57",
      "content": "<p>[quote=Jules;19094]</p>\r\n<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>\r\n<p>[/quote]This flight was indeed diverted. I'm about to hop on a plane, but I'll check the remainder of the flights in the public leaderboard set after I land, and then post information on any other potential diverted flights.</p>\r\n<p>There will be no revision of the public leaderboard set. Instead, I'll provide the actual arrival times for the diverted flight(s), and you will be free to use those in the corresponding rows for your public leaderboard submissions (though not in estimating\r\n the arrival times of any other flights in the set).</p>\r\n<p>We will ensure that there are no diverted flights in the final evaluation set.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19123",
      "postDate": "01/09/2013 08:15:42",
      "content": "<p>[quote=Ben Hamner;19107]</p>\r\n<p>[quote=Jules;19094]</p>\r\n<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>\r\n<p>[/quote]This flight was indeed diverted. I'm about to hop on a plane, but I'll check the remainder of the flights in the public leaderboard set after I land, and then post information on any other potential diverted flights.</p>\r\n<p>There will be no revision of the public leaderboard set. Instead, I'll provide the actual arrival times for the diverted flight(s), and you will be free to use those in the corresponding rows for your public leaderboard submissions (though not in estimating\r\n the arrival times of any other flights in the set).</p>\r\n<p>We will ensure that there are no diverted flights in the final evaluation set.</p>\r\n<p>[/quote]</p>\r\n<p>Need a prediction as to when your plane will be landing? :-)</p>\r\n<p>Thanks for your response on the diverted flights.</p>\r\n<p>While we are on the subject of redirected/diverted flights, could you clarify what will be done with the redirected flights.(redirected but not diverted)</p>\r\n<p>According to this link:</p>\r\n<p>https://helpdesk.flightstats.com/entries/20002503-what-are-the-flight-status-codes</p>\r\n<p>A flight is only called diverted when the plane actually lands on another airport. A flight is called redirected when it gets another destination but hasn't landed at this new destination.</p>\r\n<p>A flight can be redirected but not diverted, when a flight is redirected in mid air to another airport but is then changed back again to its original destination.</p>\r\n<p>I don't know if there are any examples of this in the test set, but there is at least one in the trainingset:</p>\r\n<p>flight_history_id=280480834 on 15th of November 2012.</p>\r\n<p>Other data (a.o. asdiposition) suggests this flight has landed around 2012-11-15 18:25 utc but the actual_runway_arrival is set to 2012-11-16 01:23:00 utc</p>\r\n<p>I suspect the actual_runway_arrival and actual_gate_arrival times for redirected flights are not very accurate, so trying to predict them would be like making a system that can predict the faults/faulty procedures in another system, I don't think that would\r\n help GE much.</p>\r\n<p>So, as is maybe clear from the above, I suggest to also leave out the redirected flights.</p>\r\n<p>Thanks again,</p>\r\n<p>Jules</p>\r\n<p>&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19158",
      "postDate": "01/10/2013 17:29:48",
      "content": "<p>These are the diverted flights I found in the test set:</p>\r\n<pre>281453856,2012-11-26 05:19:39.373000-08:00,STATUS-Diverted,&quot;AGD- New=11/26/12 07:55, STATUS- Old=R New=D&quot;<br><br>281717808,2012-11-29 09:38:35.220000-08:00,STATUS-Diverted,&quot;ARA- New=11/29/12 11:30, STATUS- Old=R New=D&quot;<br><br>281929782,2012-12-02 09:27:50.294000-08:00,STATUS-Diverted,&quot;EGA- Old=12/02/12 11:39 New=12/02/12 11:04, STATUS- Old=R New=D&quot;<br><br>282427060,2012-12-07 07:12:43.072000-08:00,STATUS-Diverted,&quot;ARA- New=12/07/12 09:04, STATUS- Old=A New=D&quot;</pre>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19286",
      "postDate": "01/14/2013 11:19:41",
      "content": "<p>Can we have a definitive answer about redirected flights?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19298",
      "postDate": "01/14/2013 16:51:57",
      "content": "<p>Will the github code be updated to filter out diverted flights?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19303",
      "postDate": "01/14/2013 21:00:05",
      "content": "<p><span>Ben,</span></p>\r\n<p>When are you going to provide&nbsp;the actual arrival times for the diverted flight(s) in PublicLeaderBoard ?</p>\r\n<p>Thanks,</p>\r\n<p>&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19512",
      "postDate": "01/19/2013 19:47:11",
      "content": "<p>Maybe Ben has not answered yet because the plane he was about to hop on was diverted or redirected. Wouldn't that be ironic.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19599",
      "postDate": "01/22/2013 19:52:32",
      "content": "<p>yes, providing the answers for such flights in publicleaderboard and excluding from the final set will be great</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19603",
      "postDate": "01/22/2013 22:33:40",
      "content": "<p>There were 24 flights in the public leaderboard set that should have been excluded due to being redirected or diverted. These are now posted in redirected_or_diverted_flights_solution.csv on the data page. You may manually replace your predictions for these\r\n 24 flights with their true values when you make submissions, and you will not be asked to make predictions for redirected or diverted flights in the final evaluation set.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19631",
      "postDate": "01/23/2013 18:30:53",
      "content": "<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19632",
      "postDate": "01/23/2013 18:38:52",
      "content": "<p>[quote=LR;19631]</p>\r\n<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>\r\n<p>[/quote]Anything that had a STATUS-Diverted or STATUS-Redirected appear in flighthistoryevents.csv will not be present in the final evaluation set.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19639",
      "postDate": "01/23/2013 20:16:44",
      "content": "<p>by replacing you get a &#43;ve score of 0.04 - not a big deal</p>\r\n<p>Use below code to remove</p>\r\n<h1>remove flights that have a status of redirected or diverted</h1>\r\n<pre><code>bad_flights &lt;- flighthistoryevents [(flighthistoryevents$event == &quot;STATUS-Diverted&quot; ) | (flighthistoryevents$event == &quot;STATUS-Redirected&quot; ),]\nflighthistory &lt;- flighthistory[!(flighthistory$flight_history_id %in% bad_flights$flight_history_id),]\n</code></pre>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "19642",
      "postDate": "01/23/2013 21:26:37",
      "content": "<p>[quote=Ben Hamner;19632]</p>\r\n<p>[quote=LR;19631]</p>\r\n<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>\r\n<p>[/quote]Anything that had a STATUS-Diverted or STATUS-Redirected appear in flighthistoryevents.csv will not be present in the final evaluation set.</p>\r\n<p>[/quote]</p>\r\n<p>&nbsp;</p>\r\n<p>How did you get 25 flights for the whole leaderboard set then? I am getting 85 for 2012-12-05 only:</p>\r\n<p>&nbsp;</p>\r\n<blockquote>\r\n<p>flighthistoryid date_time_recorded event<br>\r\n5489 282460802 2012-12-04 17:16:00.633000-08:00 STATUS-Redirected<br>\r\n6062 282464511 2012-12-05 11:48:54.297000-08:00 STATUS-Diverted<br>\r\n6070 282464511 2012-12-05 09:31:39.396000-08:00 STATUS-Redirected<br>\r\n6074 282464511 2012-12-05 06:36:55.220000-08:00 STATUS-Redirected<br>\r\n8046 282238329 2012-12-05 12:36:52.053000-08:00 STATUS-Redirected<br>\r\n14495 282465392 2012-12-05 12:06:31.619000-08:00 STATUS-Diverted<br>\r\n14504 282465392 2012-12-05 08:32:42.866000-08:00 STATUS-Redirected<br>\r\n14549 282463110 2012-12-05 01:45:09.960000-08:00 STATUS-Redirected<br>\r\n16284 282465340 2012-12-05 12:05:42.843000-08:00 STATUS-Redirected<br>\r\n20656 282463916 2012-12-05 08:56:58.597000-08:00 STATUS-Diverted<br>\r\n20658 282463916 2012-12-05 08:51:45.488000-08:00 STATUS-Redirected<br>\r\n22916 282465030 2012-12-05 08:59:05.983000-08:00 STATUS-Diverted<br>\r\n22920 282465030 2012-12-05 08:11:53.672000-08:00 STATUS-Redirected<br>\r\n24467 282463273 2012-12-05 06:56:44.900000-08:00 STATUS-Diverted</p>\r\n</blockquote>\r\n<p>&nbsp;</p>\r\n<p>&nbsp;[EDIT:] Ah ok, I got it. You gave only those in the air. ... &nbsp;Sorry for the noise.&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 18902,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "01/02/2013 17:04:44",
      "content": "<p>[quote=Dave;18892]</p>\r\n<p>What are we supposed to do with redirected/diverted&nbsp;flights?</p>\r\n<p>The data for them seems to be a mess.</p>\r\n<p>For example, check out flight_history_id=281435994 on 11/26.</p>\r\n<p>From what I can tell, it was going from LGA to MKE, but was diverted to ROC. It stayed there for about an hour, and then went to MKE.</p>\r\n<p>The data in flighthistory.csv is screwy; the actual_gate_arrival is for MKE, but the actual_runway_arrival is for ROC (and off by an hour due to a timezone mixup, I think).</p>\r\n<p>I couldn't quickly find an example, but I assume it is possible for the flight to be diverted and never continue to the original destination airport.</p>\r\n<p>Given all this, a few specific questions:</p>\r\n<p>1) Are you intending to keep diverted flights in the test set?</p>\r\n<p>2) If so, are we supposed to be predicting the arrival times for the airport it was diverted to, or the original destination airport?</p>\r\n<p>3) Have you considered capping the error for a single flight when computing the score? From a practical standpoint, there isn't a ton of difference between being off by 2 hours or 3 hours; both are terrible, but the rmse will be very different.</p>\r\n<p>&nbsp;-Dave</p>\r\n<p>[/quote]There shouldn't be any diverted flights in the public leaderboard set that you're submitting predictions on, and you won't be asked to predict any diverted flights in the final evaluation set.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 18907,
      "author_name": "dave32",
      "author_url": "",
      "post_date": "01/02/2013 17:32:01",
      "content": "<p>I think flight_history_id=281453856 on 11/26 is in the test set and was diverted.</p>\r\n<p>281453856,2012-11-26 05:19:39.373000-08:00,STATUS-Diverted,&quot;AGD- New=11/26/12 07:55, STATUS- Old=R New=D&quot;</p>\r\n<p>&nbsp; Dave</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19094,
      "author_name": "julesvanligtenberg",
      "author_url": "",
      "post_date": "01/08/2013 15:22:11",
      "content": "<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19107,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "01/08/2013 19:27:57",
      "content": "<p>[quote=Jules;19094]</p>\r\n<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>\r\n<p>[/quote]This flight was indeed diverted. I'm about to hop on a plane, but I'll check the remainder of the flights in the public leaderboard set after I land, and then post information on any other potential diverted flights.</p>\r\n<p>There will be no revision of the public leaderboard set. Instead, I'll provide the actual arrival times for the diverted flight(s), and you will be free to use those in the corresponding rows for your public leaderboard submissions (though not in estimating\r\n the arrival times of any other flights in the set).</p>\r\n<p>We will ensure that there are no diverted flights in the final evaluation set.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19123,
      "author_name": "julesvanligtenberg",
      "author_url": "",
      "post_date": "01/09/2013 08:15:42",
      "content": "<p>[quote=Ben Hamner;19107]</p>\r\n<p>[quote=Jules;19094]</p>\r\n<p>Ben,</p>\r\n<p>I agree with Dave.<br>\r\nflight<em>history</em>id=281453856 seems diverted and in the test set.<br>\r\nWill there be a revision of the test set?</p>\r\n<p>Jules</p>\r\n<p>[/quote]This flight was indeed diverted. I'm about to hop on a plane, but I'll check the remainder of the flights in the public leaderboard set after I land, and then post information on any other potential diverted flights.</p>\r\n<p>There will be no revision of the public leaderboard set. Instead, I'll provide the actual arrival times for the diverted flight(s), and you will be free to use those in the corresponding rows for your public leaderboard submissions (though not in estimating\r\n the arrival times of any other flights in the set).</p>\r\n<p>We will ensure that there are no diverted flights in the final evaluation set.</p>\r\n<p>[/quote]</p>\r\n<p>Need a prediction as to when your plane will be landing? :-)</p>\r\n<p>Thanks for your response on the diverted flights.</p>\r\n<p>While we are on the subject of redirected/diverted flights, could you clarify what will be done with the redirected flights.(redirected but not diverted)</p>\r\n<p>According to this link:</p>\r\n<p>https://helpdesk.flightstats.com/entries/20002503-what-are-the-flight-status-codes</p>\r\n<p>A flight is only called diverted when the plane actually lands on another airport. A flight is called redirected when it gets another destination but hasn't landed at this new destination.</p>\r\n<p>A flight can be redirected but not diverted, when a flight is redirected in mid air to another airport but is then changed back again to its original destination.</p>\r\n<p>I don't know if there are any examples of this in the test set, but there is at least one in the trainingset:</p>\r\n<p>flight_history_id=280480834 on 15th of November 2012.</p>\r\n<p>Other data (a.o. asdiposition) suggests this flight has landed around 2012-11-15 18:25 utc but the actual_runway_arrival is set to 2012-11-16 01:23:00 utc</p>\r\n<p>I suspect the actual_runway_arrival and actual_gate_arrival times for redirected flights are not very accurate, so trying to predict them would be like making a system that can predict the faults/faulty procedures in another system, I don't think that would\r\n help GE much.</p>\r\n<p>So, as is maybe clear from the above, I suggest to also leave out the redirected flights.</p>\r\n<p>Thanks again,</p>\r\n<p>Jules</p>\r\n<p>&nbsp;</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19158,
      "author_name": "budiansky",
      "author_url": "",
      "post_date": "01/10/2013 17:29:48",
      "content": "<p>These are the diverted flights I found in the test set:</p>\r\n<pre>281453856,2012-11-26 05:19:39.373000-08:00,STATUS-Diverted,&quot;AGD- New=11/26/12 07:55, STATUS- Old=R New=D&quot;<br><br>281717808,2012-11-29 09:38:35.220000-08:00,STATUS-Diverted,&quot;ARA- New=11/29/12 11:30, STATUS- Old=R New=D&quot;<br><br>281929782,2012-12-02 09:27:50.294000-08:00,STATUS-Diverted,&quot;EGA- Old=12/02/12 11:39 New=12/02/12 11:04, STATUS- Old=R New=D&quot;<br><br>282427060,2012-12-07 07:12:43.072000-08:00,STATUS-Diverted,&quot;ARA- New=12/07/12 09:04, STATUS- Old=A New=D&quot;</pre>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19286,
      "author_name": "elloco",
      "author_url": "",
      "post_date": "01/14/2013 11:19:41",
      "content": "<p>Can we have a definitive answer about redirected flights?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19298,
      "author_name": "jwkvam",
      "author_url": "",
      "post_date": "01/14/2013 16:51:57",
      "content": "<p>Will the github code be updated to filter out diverted flights?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19303,
      "author_name": "orchid",
      "author_url": "",
      "post_date": "01/14/2013 21:00:05",
      "content": "<p><span>Ben,</span></p>\r\n<p>When are you going to provide&nbsp;the actual arrival times for the diverted flight(s) in PublicLeaderBoard ?</p>\r\n<p>Thanks,</p>\r\n<p>&nbsp;</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19512,
      "author_name": "breakfastpirate",
      "author_url": "",
      "post_date": "01/19/2013 19:47:11",
      "content": "<p>Maybe Ben has not answered yet because the plane he was about to hop on was diverted or redirected. Wouldn't that be ironic.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19599,
      "author_name": "rkirana",
      "author_url": "",
      "post_date": "01/22/2013 19:52:32",
      "content": "<p>yes, providing the answers for such flights in publicleaderboard and excluding from the final set will be great</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19603,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "01/22/2013 22:33:40",
      "content": "<p>There were 24 flights in the public leaderboard set that should have been excluded due to being redirected or diverted. These are now posted in redirected_or_diverted_flights_solution.csv on the data page. You may manually replace your predictions for these\r\n 24 flights with their true values when you make submissions, and you will not be asked to make predictions for redirected or diverted flights in the final evaluation set.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19631,
      "author_name": "songgc",
      "author_url": "",
      "post_date": "01/23/2013 18:30:53",
      "content": "<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19632,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "01/23/2013 18:38:52",
      "content": "<p>[quote=LR;19631]</p>\r\n<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>\r\n<p>[/quote]Anything that had a STATUS-Diverted or STATUS-Redirected appear in flighthistoryevents.csv will not be present in the final evaluation set.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19639,
      "author_name": "rkirana",
      "author_url": "",
      "post_date": "01/23/2013 20:16:44",
      "content": "<p>by replacing you get a &#43;ve score of 0.04 - not a big deal</p>\r\n<p>Use below code to remove</p>\r\n<h1>remove flights that have a status of redirected or diverted</h1>\r\n<pre><code>bad_flights &lt;- flighthistoryevents [(flighthistoryevents$event == &quot;STATUS-Diverted&quot; ) | (flighthistoryevents$event == &quot;STATUS-Redirected&quot; ),]\nflighthistory &lt;- flighthistory[!(flighthistory$flight_history_id %in% bad_flights$flight_history_id),]\n</code></pre>",
      "votes": null,
      "replies": []
    },
    {
      "id": 19642,
      "author_name": "vspinu",
      "author_url": "",
      "post_date": "01/23/2013 21:26:37",
      "content": "<p>[quote=Ben Hamner;19632]</p>\r\n<p>[quote=LR;19631]</p>\r\n<p>Ben,</p>\r\n<p><span style=\"line-height:1.4em\">We need to clean redirected/diverted flights in the training set as well. Can you specify the rules that indentify these flights? or publish sample codes?</span></p>\r\n<p><span style=\"line-height:1.4em\">Thanks!</span></p>\r\n<p>[/quote]Anything that had a STATUS-Diverted or STATUS-Redirected appear in flighthistoryevents.csv will not be present in the final evaluation set.</p>\r\n<p>[/quote]</p>\r\n<p>&nbsp;</p>\r\n<p>How did you get 25 flights for the whole leaderboard set then? I am getting 85 for 2012-12-05 only:</p>\r\n<p>&nbsp;</p>\r\n<blockquote>\r\n<p>flighthistoryid date_time_recorded event<br>\r\n5489 282460802 2012-12-04 17:16:00.633000-08:00 STATUS-Redirected<br>\r\n6062 282464511 2012-12-05 11:48:54.297000-08:00 STATUS-Diverted<br>\r\n6070 282464511 2012-12-05 09:31:39.396000-08:00 STATUS-Redirected<br>\r\n6074 282464511 2012-12-05 06:36:55.220000-08:00 STATUS-Redirected<br>\r\n8046 282238329 2012-12-05 12:36:52.053000-08:00 STATUS-Redirected<br>\r\n14495 282465392 2012-12-05 12:06:31.619000-08:00 STATUS-Diverted<br>\r\n14504 282465392 2012-12-05 08:32:42.866000-08:00 STATUS-Redirected<br>\r\n14549 282463110 2012-12-05 01:45:09.960000-08:00 STATUS-Redirected<br>\r\n16284 282465340 2012-12-05 12:05:42.843000-08:00 STATUS-Redirected<br>\r\n20656 282463916 2012-12-05 08:56:58.597000-08:00 STATUS-Diverted<br>\r\n20658 282463916 2012-12-05 08:51:45.488000-08:00 STATUS-Redirected<br>\r\n22916 282465030 2012-12-05 08:59:05.983000-08:00 STATUS-Diverted<br>\r\n22920 282465030 2012-12-05 08:11:53.672000-08:00 STATUS-Redirected<br>\r\n24467 282463273 2012-12-05 06:56:44.900000-08:00 STATUS-Diverted</p>\r\n</blockquote>\r\n<p>&nbsp;</p>\r\n<p>&nbsp;[EDIT:] Ah ok, I got it. You gave only those in the air. ... &nbsp;Sorry for the noise.&nbsp;</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "18892": "",
    "18902": "",
    "18907": "",
    "19094": "",
    "19107": "",
    "19123": "",
    "19158": "",
    "19286": "",
    "19298": "",
    "19303": "",
    "19512": "",
    "19599": "",
    "19603": "",
    "19631": "",
    "19632": "",
    "19639": "",
    "19642": ""
  },
  "source": "meta"
}