{
  "id": 223181,
  "title": "Last Seen Timestamp",
  "url": "/competitions/indoor-location-navigation/discussion/223181",
  "author_name": "",
  "post_date": "2021-03-02T19:44:57.280032700Z",
  "votes": 9,
  "comment_count": 7,
  "views": 0,
  "content": "<p>Hi Kagglers! What do you think? How much value can WiFi's \"last seen timestamp\" attribute bring into our predictions. If any, how could we justify it?</p>",
  "messages": [
    {
      "id": "1224521",
      "postDate": "03/02/2021 19:44:57",
      "content": "<p>Hi Kagglers! What do you think? How much value can WiFi's \"last seen timestamp\" attribute bring into our predictions. If any, how could we justify it?</p>",
      "rawMarkdown": "Hi Kagglers! What do you think? How much value can WiFi's \"last seen timestamp\" attribute bring into our predictions. If any, how could we justify it?",
      "votes": null
    },
    {
      "id": "1224533",
      "postDate": "03/02/2021 20:10:56",
      "content": "<p><code>last_timestamp</code> is much better than <code>timestamp</code>. Please see this video for an explanation. <a href=\"https://youtu.be/xt3OzMC-XMU?t=690\" target=\"_blank\">https://youtu.be/xt3OzMC-XMU?t=690</a></p>",
      "rawMarkdown": "`last_timestamp` is much better than `timestamp`. Please see this video for an explanation. https://youtu.be/xt3OzMC-XMU?t=690",
      "votes": null
    },
    {
      "id": "1224574",
      "postDate": "03/02/2021 21:38:14",
      "content": "<p>Thank you for your reply! Does it imply that WiFi APs that appear in data can theoretically be from outside the building (they were seen before the path began)? </p>\n<p><code>sum(data.wifi[:,-1].astype(int) - data.waypoint[0][0] &lt; 0)</code> returns 40 for some path, which means that some APs were detected before the path began. Am I correct? :)</p>",
      "rawMarkdown": "Thank you for your reply! Does it imply that WiFi APs that appear in data can theoretically be from outside the building (they were seen before the path began)? \n\n`sum(data.wifi[:,-1].astype(int) - data.waypoint[0][0] < 0)` returns 40 for some path, which means that some APs were detected before the path began. Am I correct? :)",
      "votes": null
    },
    {
      "id": "1224579",
      "postDate": "03/02/2021 21:43:57",
      "content": "<p>Yeah, I think so. 👌</p>",
      "rawMarkdown": "Yeah, I think so. 👌",
      "votes": null
    },
    {
      "id": "1224591",
      "postDate": "03/02/2021 22:01:22",
      "content": "<p>I agree. I ignore any reading that occurred more than 5 seconds before the first waypoint, which appears to be pretty conservative</p>",
      "rawMarkdown": "I agree. I ignore any reading that occurred more than 5 seconds before the first waypoint, which appears to be pretty conservative",
      "votes": null
    },
    {
      "id": "1237429",
      "postDate": "03/14/2021 06:16:54",
      "content": "<p><a href=\"https://www.kaggle.com/jiweiliu\" target=\"_blank\">@jiweiliu</a> Thanks for sharing. I am curious about how to use last_timestamp in testset, timestamp and last_timestamp are not in the same dimension?</p>",
      "rawMarkdown": "jiweiliu Thanks for sharing. I am curious about how to use last_timestamp in testset, timestamp and last_timestamp are not in the same dimension?",
      "votes": null
    },
    {
      "id": "1239467",
      "postDate": "03/15/2021 18:34:15",
      "content": "<p>Please take a look at this <a href=\"https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask\" target=\"_blank\">https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask</a></p>",
      "rawMarkdown": "Please take a look at this https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask",
      "votes": null
    },
    {
      "id": "1247800",
      "postDate": "03/22/2021 04:29:37",
      "content": "<p><a href=\"https://www.kaggle.com/jiweiliu\" target=\"_blank\">@jiweiliu</a> Thanks for sharing!</p>",
      "rawMarkdown": "jiweiliu Thanks for sharing!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1224533,
      "author_name": "jiweiliu",
      "author_url": "",
      "post_date": "03/02/2021 20:10:56",
      "content": "<p><code>last_timestamp</code> is much better than <code>timestamp</code>. Please see this video for an explanation. <a href=\"https://youtu.be/xt3OzMC-XMU?t=690\" target=\"_blank\">https://youtu.be/xt3OzMC-XMU?t=690</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 1224574,
          "author_name": "olaf2000",
          "author_url": "",
          "post_date": "03/02/2021 21:38:14",
          "content": "<p>Thank you for your reply! Does it imply that WiFi APs that appear in data can theoretically be from outside the building (they were seen before the path began)? </p>\n<p><code>sum(data.wifi[:,-1].astype(int) - data.waypoint[0][0] &lt; 0)</code> returns 40 for some path, which means that some APs were detected before the path began. Am I correct? :)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1224579,
          "author_name": "jiweiliu",
          "author_url": "",
          "post_date": "03/02/2021 21:43:57",
          "content": "<p>Yeah, I think so. 👌</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1224591,
          "author_name": "nigelhenry",
          "author_url": "",
          "post_date": "03/02/2021 22:01:22",
          "content": "<p>I agree. I ignore any reading that occurred more than 5 seconds before the first waypoint, which appears to be pretty conservative</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1237429,
          "author_name": "chenxin1991",
          "author_url": "",
          "post_date": "03/14/2021 06:16:54",
          "content": "<p><a href=\"https://www.kaggle.com/jiweiliu\" target=\"_blank\">@jiweiliu</a> Thanks for sharing. I am curious about how to use last_timestamp in testset, timestamp and last_timestamp are not in the same dimension?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1239467,
          "author_name": "jiweiliu",
          "author_url": "",
          "post_date": "03/15/2021 18:34:15",
          "content": "<p>Please take a look at this <a href=\"https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask\" target=\"_blank\">https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1247800,
          "author_name": "chenxin1991",
          "author_url": "",
          "post_date": "03/22/2021 04:29:37",
          "content": "<p><a href=\"https://www.kaggle.com/jiweiliu\" target=\"_blank\">@jiweiliu</a> Thanks for sharing!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1224521": "Hi Kagglers! What do you think? How much value can WiFi's \"last seen timestamp\" attribute bring into our predictions. If any, how could we justify it?",
    "1224533": "`last_timestamp` is much better than `timestamp`. Please see this video for an explanation. https://youtu.be/xt3OzMC-XMU?t=690",
    "1224574": "Thank you for your reply! Does it imply that WiFi APs that appear in data can theoretically be from outside the building (they were seen before the path began)? \n\n`sum(data.wifi[:,-1].astype(int) - data.waypoint[0][0] < 0)` returns 40 for some path, which means that some APs were detected before the path began. Am I correct? :)",
    "1224579": "Yeah, I think so. 👌",
    "1224591": "I agree. I ignore any reading that occurred more than 5 seconds before the first waypoint, which appears to be pretty conservative",
    "1237429": "jiweiliu Thanks for sharing. I am curious about how to use last_timestamp in testset, timestamp and last_timestamp are not in the same dimension?",
    "1239467": "Please take a look at this https://www.kaggle.com/jiweiliu/fix-the-timestamps-of-test-data-using-dask",
    "1247800": "jiweiliu Thanks for sharing!"
  },
  "source": "meta"
}