{
  "id": 76579,
  "title": "Category 2 (region) - State ",
  "url": "/competitions/elo-merchant-category-recommendation/discussion/76579",
  "author_name": "",
  "post_date": "2019-01-04T11:19:03.768629500Z",
  "votes": 65,
  "comment_count": 7,
  "views": 0,
  "content": "<p>I believe that some of you already verified that <strong>\"category 2\"</strong> is completly correlated with <strong>\"state_id\"</strong> on \"historical_transactions.csv\" and \"new_merchant_transactions.csv\" but what this in fact means ? Well, here in Brazil the states are grouped into <strong>FIVE</strong> distinct \"regions\" as yout can see in the picture below :  <img src=\"https://brasilescola.uol.com.br/upload/e/estadosbrasileiros(1).jpg\" alt=\"enter image description here\"></p>\n\n<p>Guess what! Category 2 has <strong>FIVE</strong> possible values and if we group them with state id counting the number of unique cities we obtain this table:</p>\n\n<p><img src=\"https://i.ibb.co/XFTVpSk/table-region.png\" alt=\"enter image description here\"></p>\n\n<p>Now, region <strong>centro-oeste</strong> and region <strong>sudeste</strong> have four states so, how to I guessed the correct classification ?  Well, region \"centro-oeste\" has a state called \"distrito federal\" that has just one city on it, so if you check the number of unique cities on state 22 it has only one :D</p>\n\n<p>I hope that this can be helpfull in some way.</p>\n\n<p>Another statement: I believe \"category_1\" is \"Y\" for international spending and \"N\" for otherwise.</p>",
  "messages": [
    {
      "id": "450133",
      "postDate": "01/04/2019 11:19:03",
      "content": "<p>I believe that some of you already verified that <strong>\"category 2\"</strong> is completly correlated with <strong>\"state_id\"</strong> on \"historical_transactions.csv\" and \"new_merchant_transactions.csv\" but what this in fact means ? Well, here in Brazil the states are grouped into <strong>FIVE</strong> distinct \"regions\" as yout can see in the picture below :  <img src=\"https://brasilescola.uol.com.br/upload/e/estadosbrasileiros(1).jpg\" alt=\"enter image description here\"></p>\n\n<p>Guess what! Category 2 has <strong>FIVE</strong> possible values and if we group them with state id counting the number of unique cities we obtain this table:</p>\n\n<p><img src=\"https://i.ibb.co/XFTVpSk/table-region.png\" alt=\"enter image description here\"></p>\n\n<p>Now, region <strong>centro-oeste</strong> and region <strong>sudeste</strong> have four states so, how to I guessed the correct classification ?  Well, region \"centro-oeste\" has a state called \"distrito federal\" that has just one city on it, so if you check the number of unique cities on state 22 it has only one :D</p>\n\n<p>I hope that this can be helpfull in some way.</p>\n\n<p>Another statement: I believe \"category_1\" is \"Y\" for international spending and \"N\" for otherwise.</p>",
      "rawMarkdown": "I believe that some of you already verified that **\"category 2\"** is completly correlated with **\"state_id\"** on \"historical_transactions.csv\" and \"new_merchant_transactions.csv\" but what this in fact means ? Well, here in Brazil the states are grouped into **FIVE** distinct \"regions\" as yout can see in the picture below :  ![enter image description here][1]\n\nGuess what! Category 2 has **FIVE** possible values and if we group them with state id counting the number of unique cities we obtain this table:\n\n![enter image description here][2]\n\nNow, region **centro-oeste** and region **sudeste** have four states so, how to I guessed the correct classification ?  Well, region \"centro-oeste\" has a state called \"distrito federal\" that has just one city on it, so if you check the number of unique cities on state 22 it has only one :D\n\nI hope that this can be helpfull in some way.\n\nAnother statement: I believe \"category_1\" is \"Y\" for international spending and \"N\" for otherwise.\n\n  [1]: https://brasilescola.uol.com.br/upload/e/estadosbrasileiros(1).jpg\n  [2]: https://i.ibb.co/XFTVpSk/table-region.png",
      "votes": null
    },
    {
      "id": "450366",
      "postDate": "01/04/2019 19:28:56",
      "content": "<p>Nice work!\nI guess your country of origin helps a little bit!</p>",
      "rawMarkdown": "Nice work!\nI guess your country of origin helps a little bit!",
      "votes": null
    },
    {
      "id": "450433",
      "postDate": "01/04/2019 23:52:30",
      "content": "<p>brilliant!</p>",
      "rawMarkdown": "brilliant!",
      "votes": null
    },
    {
      "id": "451733",
      "postDate": "01/07/2019 15:27:27",
      "content": "<p>While this seems to be the case, I decided to check the correlation between each of the variables before deciding which variables to keep, and which variables to drop. We all could check it out</p>",
      "rawMarkdown": "While this seems to be the case, I decided to check the correlation between each of the variables before deciding which variables to keep, and which variables to drop. We all could check it out",
      "votes": null
    },
    {
      "id": "452655",
      "postDate": "01/09/2019 02:09:47",
      "content": "<p>good work :)</p>",
      "rawMarkdown": "good work :)",
      "votes": null
    },
    {
      "id": "453417",
      "postDate": "01/10/2019 07:04:31",
      "content": "<p>Thanks for notifying.\nUsing these intel, maybe we can add some geographical features like latitude.</p>",
      "rawMarkdown": "Thanks for notifying.\nUsing these intel, maybe we can add some geographical features like latitude.",
      "votes": null
    },
    {
      "id": "456472",
      "postDate": "01/15/2019 21:52:12",
      "content": "<p>Muito bom grande Angelo!\nVo ver se me dedico agora tambem!\nGrande Abraco.</p>",
      "rawMarkdown": "Muito bom grande Angelo!\nVo ver se me dedico agora tambem!\nGrande Abraco.",
      "votes": null
    },
    {
      "id": "465748",
      "postDate": "02/03/2019 22:39:26",
      "content": "<p>I wonder if the customers' state id has any effect on their loyalty scores though</p>",
      "rawMarkdown": "I wonder if the customers' state id has any effect on their loyalty scores though",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 450366,
      "author_name": "bezul2",
      "author_url": "",
      "post_date": "01/04/2019 19:28:56",
      "content": "<p>Nice work!\nI guess your country of origin helps a little bit!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 450433,
      "author_name": "liwenhuang",
      "author_url": "",
      "post_date": "01/04/2019 23:52:30",
      "content": "<p>brilliant!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 451733,
      "author_name": "sikolia",
      "author_url": "",
      "post_date": "01/07/2019 15:27:27",
      "content": "<p>While this seems to be the case, I decided to check the correlation between each of the variables before deciding which variables to keep, and which variables to drop. We all could check it out</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 452655,
      "author_name": "kalyankkr",
      "author_url": "",
      "post_date": "01/09/2019 02:09:47",
      "content": "<p>good work :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 453417,
      "author_name": "mutsumichieda",
      "author_url": "",
      "post_date": "01/10/2019 07:04:31",
      "content": "<p>Thanks for notifying.\nUsing these intel, maybe we can add some geographical features like latitude.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 456472,
      "author_name": "paolorampazzo",
      "author_url": "",
      "post_date": "01/15/2019 21:52:12",
      "content": "<p>Muito bom grande Angelo!\nVo ver se me dedico agora tambem!\nGrande Abraco.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 465748,
      "author_name": "liwenhuang",
      "author_url": "",
      "post_date": "02/03/2019 22:39:26",
      "content": "<p>I wonder if the customers' state id has any effect on their loyalty scores though</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "450133": "I believe that some of you already verified that **\"category 2\"** is completly correlated with **\"state_id\"** on \"historical_transactions.csv\" and \"new_merchant_transactions.csv\" but what this in fact means ? Well, here in Brazil the states are grouped into **FIVE** distinct \"regions\" as yout can see in the picture below :  ![enter image description here][1]\n\nGuess what! Category 2 has **FIVE** possible values and if we group them with state id counting the number of unique cities we obtain this table:\n\n![enter image description here][2]\n\nNow, region **centro-oeste** and region **sudeste** have four states so, how to I guessed the correct classification ?  Well, region \"centro-oeste\" has a state called \"distrito federal\" that has just one city on it, so if you check the number of unique cities on state 22 it has only one :D\n\nI hope that this can be helpfull in some way.\n\nAnother statement: I believe \"category_1\" is \"Y\" for international spending and \"N\" for otherwise.\n\n  [1]: https://brasilescola.uol.com.br/upload/e/estadosbrasileiros(1).jpg\n  [2]: https://i.ibb.co/XFTVpSk/table-region.png",
    "450366": "Nice work!\nI guess your country of origin helps a little bit!",
    "450433": "brilliant!",
    "451733": "While this seems to be the case, I decided to check the correlation between each of the variables before deciding which variables to keep, and which variables to drop. We all could check it out",
    "452655": "good work :)",
    "453417": "Thanks for notifying.\nUsing these intel, maybe we can add some geographical features like latitude.",
    "456472": "Muito bom grande Angelo!\nVo ver se me dedico agora tambem!\nGrande Abraco.",
    "465748": "I wonder if the customers' state id has any effect on their loyalty scores though"
  },
  "source": "meta"
}