{
  "id": 51767,
  "title": "Description of Features",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/51767",
  "author_name": "",
  "post_date": "2018-03-12T23:06:53.533409300Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi there: \nI'm a little bit confusing about the features in this data set: \n1. 'ip: ip address of click'\nHow to understand the 'ip address of click', does it mean that ip address of the network or device ?\n2. 'Channel: channel id of mobile ad publisher'\nWhat does this 'channel id' properly ? \n3. 'attributed_time: if user download the app for after clicking an ad, this is the time of the app download'\nDoes it mean the timestamp that downloading starts after clicking ?</p>",
  "messages": [
    {
      "id": "294981",
      "postDate": "03/12/2018 23:06:53",
      "content": "<p>Hi there: \nI'm a little bit confusing about the features in this data set: \n1. 'ip: ip address of click'\nHow to understand the 'ip address of click', does it mean that ip address of the network or device ?\n2. 'Channel: channel id of mobile ad publisher'\nWhat does this 'channel id' properly ? \n3. 'attributed_time: if user download the app for after clicking an ad, this is the time of the app download'\nDoes it mean the timestamp that downloading starts after clicking ?</p>",
      "rawMarkdown": "Hi there: \nI'm a little bit confusing about the features in this data set: \n1. 'ip: ip address of click'\nHow to understand the 'ip address of click', does it mean that ip address of the network or device ?\n2. 'Channel: channel id of mobile ad publisher'\nWhat does this 'channel id' properly ? \n3. 'attributed_time: if user download the app for after clicking an ad, this is the time of the app download'\nDoes it mean the timestamp that downloading starts after clicking ?",
      "votes": null
    },
    {
      "id": "295181",
      "postDate": "03/13/2018 07:48:08",
      "content": "<p>My understanding on the questions is: 1) IP of the user that clicked (network in that sense I guess, multiple users can have the same IP and at the same time, a given user can change IPs throughout time) 2) imagine the channels like - facebook, google ads, RTB ads on \"random\" websites, etc. 3) the timestamp when the downloading starts (and supposedly the user has paid for it:) )</p>",
      "rawMarkdown": "My understanding on the questions is: 1) IP of the user that clicked (network in that sense I guess, multiple users can have the same IP and at the same time, a given user can change IPs throughout time) 2) imagine the channels like - facebook, google ads, RTB ads on \"random\" websites, etc. 3) the timestamp when the downloading starts (and supposedly the user has paid for it:) )",
      "votes": null
    },
    {
      "id": "301790",
      "postDate": "03/23/2018 08:56:32",
      "content": "<p>All of these things are from the point of view of online ad impressions served up to users:</p>\n\n<ol>\n<li><code>channel id</code> in online advertising would be an id which could mean Weibo, Wechat, Baidu, Alibaba, Tencent etc.\nNot that it matters for this competition, but here's a <a href=\"https://www.entrepreneur.com/article/239499\">background article</a>. (Also, not relevant here, but the advertiser could also have multiple distinct campaign ids, e.g. display ad, paid search listing etc.) The individual channel id's will have different CTR rates and fraud rates.</li>\n<li>Is  <code>attributed_time</code> the start time of user download? Maybe, I don't know. If you look at clickstreams with multiple clicks, you should be able to figure that out.</li>\n</ol>",
      "rawMarkdown": "All of these things are from the point of view of online ad impressions served up to users:\n\n2. `channel id` in online advertising would be an id which could mean Weibo, Wechat, Baidu, Alibaba, Tencent etc.\n Not that it matters for this competition, but here's a [background article][1]. (Also, not relevant here, but the advertiser could also have multiple distinct campaign ids, e.g. display ad, paid search listing etc.) The individual channel id's will have different CTR rates and fraud rates.\n3. Is  `attributed_time` the start time of user download? Maybe, I don't know. If you look at clickstreams with multiple clicks, you should be able to figure that out.\n\n  [1]: https://www.entrepreneur.com/article/239499",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 295181,
      "author_name": "asparuhhristov",
      "author_url": "",
      "post_date": "03/13/2018 07:48:08",
      "content": "<p>My understanding on the questions is: 1) IP of the user that clicked (network in that sense I guess, multiple users can have the same IP and at the same time, a given user can change IPs throughout time) 2) imagine the channels like - facebook, google ads, RTB ads on \"random\" websites, etc. 3) the timestamp when the downloading starts (and supposedly the user has paid for it:) )</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 301790,
      "author_name": "smcinerney",
      "author_url": "",
      "post_date": "03/23/2018 08:56:32",
      "content": "<p>All of these things are from the point of view of online ad impressions served up to users:</p>\n\n<ol>\n<li><code>channel id</code> in online advertising would be an id which could mean Weibo, Wechat, Baidu, Alibaba, Tencent etc.\nNot that it matters for this competition, but here's a <a href=\"https://www.entrepreneur.com/article/239499\">background article</a>. (Also, not relevant here, but the advertiser could also have multiple distinct campaign ids, e.g. display ad, paid search listing etc.) The individual channel id's will have different CTR rates and fraud rates.</li>\n<li>Is  <code>attributed_time</code> the start time of user download? Maybe, I don't know. If you look at clickstreams with multiple clicks, you should be able to figure that out.</li>\n</ol>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "294981": "Hi there: \nI'm a little bit confusing about the features in this data set: \n1. 'ip: ip address of click'\nHow to understand the 'ip address of click', does it mean that ip address of the network or device ?\n2. 'Channel: channel id of mobile ad publisher'\nWhat does this 'channel id' properly ? \n3. 'attributed_time: if user download the app for after clicking an ad, this is the time of the app download'\nDoes it mean the timestamp that downloading starts after clicking ?",
    "295181": "My understanding on the questions is: 1) IP of the user that clicked (network in that sense I guess, multiple users can have the same IP and at the same time, a given user can change IPs throughout time) 2) imagine the channels like - facebook, google ads, RTB ads on \"random\" websites, etc. 3) the timestamp when the downloading starts (and supposedly the user has paid for it:) )",
    "301790": "All of these things are from the point of view of online ad impressions served up to users:\n\n2. `channel id` in online advertising would be an id which could mean Weibo, Wechat, Baidu, Alibaba, Tencent etc.\n Not that it matters for this competition, but here's a [background article][1]. (Also, not relevant here, but the advertiser could also have multiple distinct campaign ids, e.g. display ad, paid search listing etc.) The individual channel id's will have different CTR rates and fraud rates.\n3. Is  `attributed_time` the start time of user download? Maybe, I don't know. If you look at clickstreams with multiple clicks, you should be able to figure that out.\n\n  [1]: https://www.entrepreneur.com/article/239499"
  },
  "source": "meta"
}