{
  "id": 76217,
  "title": "Month Lag Demystified",
  "url": "/competitions/elo-merchant-category-recommendation/discussion/76217",
  "author_name": "",
  "post_date": "2018-12-30T17:55:49.493424200Z",
  "votes": 34,
  "comment_count": 1,
  "views": 0,
  "content": "<p>In the Elo merchant category recommendation competition, month lag column exist in both historical_transaction &amp; new_merchant_transactions tables. </p>\n\n<p>The significance of month_lag?\n- <strong>Reference date</strong> - Last transaction date of any particular card id in historical_transaction is called reference date for that card_id\n- <strong>Definition</strong>: Month_lag transactions for any card_id, month(purchase_date) - reference date for that card_id</p>\n\n<p><em>Example -</em>\nFor card ID - C_ID_0002c7c2c1\nLast transaction date in historical_transaction = 2018-01-25 11:13:39\nThus, reference date for card ID C_ID_0002c7c2c1 = 2018-01-25\nNew merchant transaction data for this card ID has 6 transactions but all of them are done in the month 2018-02 &amp; 2018-03, which is reference_date+1 (month lag = 1), reference_date+2 (month lag = 2)</p>\n\n<p>This might help solve queries like - \n1. Why there are 12 months of data in new_merchant_transactions?\n- Based on the above information, for any card ID only 2 month of data is given. These 2 months are the ones following reference date.</p>\n\n<p>Hope this helps!</p>",
  "messages": [
    {
      "id": "447843",
      "postDate": "12/30/2018 17:55:49",
      "content": "<p>In the Elo merchant category recommendation competition, month lag column exist in both historical_transaction &amp; new_merchant_transactions tables. </p>\n\n<p>The significance of month_lag?\n- <strong>Reference date</strong> - Last transaction date of any particular card id in historical_transaction is called reference date for that card_id\n- <strong>Definition</strong>: Month_lag transactions for any card_id, month(purchase_date) - reference date for that card_id</p>\n\n<p><em>Example -</em>\nFor card ID - C_ID_0002c7c2c1\nLast transaction date in historical_transaction = 2018-01-25 11:13:39\nThus, reference date for card ID C_ID_0002c7c2c1 = 2018-01-25\nNew merchant transaction data for this card ID has 6 transactions but all of them are done in the month 2018-02 &amp; 2018-03, which is reference_date+1 (month lag = 1), reference_date+2 (month lag = 2)</p>\n\n<p>This might help solve queries like - \n1. Why there are 12 months of data in new_merchant_transactions?\n- Based on the above information, for any card ID only 2 month of data is given. These 2 months are the ones following reference date.</p>\n\n<p>Hope this helps!</p>",
      "rawMarkdown": "In the Elo merchant category recommendation competition, month lag column exist in both historical_transaction &amp; new_merchant_transactions tables. \n\nThe significance of month_lag?\n- **Reference date** - Last transaction date of any particular card id in historical_transaction is called reference date for that card_id\n- **Definition**: Month_lag transactions for any card_id, month(purchase_date) - reference date for that card_id\n\n*Example -*\nFor card ID - C_ID_0002c7c2c1\nLast transaction date in historical_transaction = 2018-01-25 11:13:39\nThus, reference date for card ID C_ID_0002c7c2c1 = 2018-01-25\nNew merchant transaction data for this card ID has 6 transactions but all of them are done in the month 2018-02 &amp; 2018-03, which is reference_date+1 (month lag = 1), reference_date+2 (month lag = 2)\n\nThis might help solve queries like - \n1. Why there are 12 months of data in new_merchant_transactions?\n- Based on the above information, for any card ID only 2 month of data is given. These 2 months are the ones following reference date.\n\nHope this helps!",
      "votes": null
    },
    {
      "id": "466531",
      "postDate": "02/05/2019 15:11:24",
      "content": "<p>Thanks for sharing! I see many folks using 2018-2-1 as a point to compute a feature called elapse_time or something like that. Do you happen to know what is the significance of that particular date?</p>",
      "rawMarkdown": "Thanks for sharing! I see many folks using 2018-2-1 as a point to compute a feature called elapse_time or something like that. Do you happen to know what is the significance of that particular date?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 466531,
      "author_name": "liwenhuang",
      "author_url": "",
      "post_date": "02/05/2019 15:11:24",
      "content": "<p>Thanks for sharing! I see many folks using 2018-2-1 as a point to compute a feature called elapse_time or something like that. Do you happen to know what is the significance of that particular date?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "447843": "In the Elo merchant category recommendation competition, month lag column exist in both historical_transaction &amp; new_merchant_transactions tables. \n\nThe significance of month_lag?\n- **Reference date** - Last transaction date of any particular card id in historical_transaction is called reference date for that card_id\n- **Definition**: Month_lag transactions for any card_id, month(purchase_date) - reference date for that card_id\n\n*Example -*\nFor card ID - C_ID_0002c7c2c1\nLast transaction date in historical_transaction = 2018-01-25 11:13:39\nThus, reference date for card ID C_ID_0002c7c2c1 = 2018-01-25\nNew merchant transaction data for this card ID has 6 transactions but all of them are done in the month 2018-02 &amp; 2018-03, which is reference_date+1 (month lag = 1), reference_date+2 (month lag = 2)\n\nThis might help solve queries like - \n1. Why there are 12 months of data in new_merchant_transactions?\n- Based on the above information, for any card ID only 2 month of data is given. These 2 months are the ones following reference date.\n\nHope this helps!",
    "466531": "Thanks for sharing! I see many folks using 2018-2-1 as a point to compute a feature called elapse_time or something like that. Do you happen to know what is the significance of that particular date?"
  },
  "source": "meta"
}