{
  "id": 13871,
  "title": "How to interpret the time data",
  "url": "/competitions/facebook-recruiting-iv-human-or-bot/discussion/13871",
  "author_name": "",
  "post_date": "2015-04-30T01:28:39.907Z",
  "votes": -1,
  "comment_count": 4,
  "views": 1537,
  "content": "<p>Hello,</p>\n<p>Are we to interpret&nbsp;the time&nbsp;data provided in the bids.csv file as a UNIX timestamp in microseconds? I noticed that the max is&nbsp;9772885210526315 and the min is&nbsp;9631916842105263 and converting these would mean we've dates between the years 2275 and 2279. I suppose it's because the data is scrambled to protect privacy. Just wanted to make sure.</p>\n<p>Best regards,</p>\n<p>Andre</p>",
  "messages": [
    {
      "id": "75541",
      "postDate": "04/30/2015 01:28:39",
      "content": "<p>Hello,</p>\n<p>Are we to interpret&nbsp;the time&nbsp;data provided in the bids.csv file as a UNIX timestamp in microseconds? I noticed that the max is&nbsp;9772885210526315 and the min is&nbsp;9631916842105263 and converting these would mean we've dates between the years 2275 and 2279. I suppose it's because the data is scrambled to protect privacy. Just wanted to make sure.</p>\n<p>Best regards,</p>\n<p>Andre</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75590",
      "postDate": "04/30/2015 09:12:06",
      "content": "<p>My guess is that we don't know how to convert the int back to datetime, but the order&nbsp;is preserved.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75716",
      "postDate": "04/30/2015 18:44:03",
      "content": "<p>I can't tell it too.</p>\n<p>All i know is, that the mean of the human bidders is a lot lower than the one of the bots.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75732",
      "postDate": "04/30/2015 19:58:03",
      "content": "<p>I disagree. In training data set (bids):</p>\n<p>&gt; mean(log(bids[bids$outcome == 0,]$time))<br>[1] -707.5612<br>&gt; mean(log(bids[bids$outcome == 1,]$time))<br>[1] -707.5589<br>&gt; var(log(bids[bids$outcome == 1,]$time))<br>[1] 9.294271e-05<br>&gt; var(log(bids[bids$outcome == 0,]$time))<br>[1] 0.000103042</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "75845",
      "postDate": "05/01/2015 02:42:47",
      "content": "<p>Hmm, this is what it looks like for me tho.</p>\n<p>In[12]: joined_train[joined_train.outcome == 0].time.mean()<br>Out[12]: 2660997069409.8838<br>In[13]: joined_train[joined_train.outcome == 1].time.mean()<br>Out[13]: 798324254165.82153</p>\n<p>In[14]: joined_train[joined_train.outcome == 0].time.mean() - joined_train[joined_train.outcome == 1].time.mean()<br>Out[14]: 1862672815244.0623</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 75590,
      "author_name": "bluesilence",
      "author_url": "",
      "post_date": "04/30/2015 09:12:06",
      "content": "<p>My guess is that we don't know how to convert the int back to datetime, but the order&nbsp;is preserved.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 75716,
      "author_name": "tomate",
      "author_url": "",
      "post_date": "04/30/2015 18:44:03",
      "content": "<p>I can't tell it too.</p>\n<p>All i know is, that the mean of the human bidders is a lot lower than the one of the bots.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 75732,
      "author_name": "lnicalo",
      "author_url": "",
      "post_date": "04/30/2015 19:58:03",
      "content": "<p>I disagree. In training data set (bids):</p>\n<p>&gt; mean(log(bids[bids$outcome == 0,]$time))<br>[1] -707.5612<br>&gt; mean(log(bids[bids$outcome == 1,]$time))<br>[1] -707.5589<br>&gt; var(log(bids[bids$outcome == 1,]$time))<br>[1] 9.294271e-05<br>&gt; var(log(bids[bids$outcome == 0,]$time))<br>[1] 0.000103042</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 75845,
      "author_name": "tomate",
      "author_url": "",
      "post_date": "05/01/2015 02:42:47",
      "content": "<p>Hmm, this is what it looks like for me tho.</p>\n<p>In[12]: joined_train[joined_train.outcome == 0].time.mean()<br>Out[12]: 2660997069409.8838<br>In[13]: joined_train[joined_train.outcome == 1].time.mean()<br>Out[13]: 798324254165.82153</p>\n<p>In[14]: joined_train[joined_train.outcome == 0].time.mean() - joined_train[joined_train.outcome == 1].time.mean()<br>Out[14]: 1862672815244.0623</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "75541": "",
    "75590": "",
    "75716": "",
    "75732": "",
    "75845": ""
  },
  "source": "meta"
}