{
  "id": 65909,
  "title": "Quick start - Python/Pandas",
  "url": "/competitions/ga-customer-revenue-prediction/discussion/65909",
  "author_name": "",
  "post_date": "2018-09-16T03:49:34.020540700Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I did a <a href=\"https://www.kaggle.com/jsaguiar/python-quick-start-clean-and-pickle-data\">script</a> to convert, clean and save the data as a pickle file. The preprocessing shouldn't throw away (or change) any usefull information.</p>\n\n<p>A quick summary:</p>\n\n<ul>\n<li>Convert json object to columns</li>\n<li>Drop constant columns (single unique value)</li>\n<li>Also drop campaignCode column (all missing except one row)</li>\n<li>Set datatype for a few numerical columns</li>\n<li>Save as pickle file</li>\n</ul>\n\n<p>To load the data: pandas.read_pickle('file_name'). If you find anything wrong please leave a comment.</p>",
  "messages": [
    {
      "id": "387983",
      "postDate": "09/16/2018 03:49:34",
      "content": "<p>I did a <a href=\"https://www.kaggle.com/jsaguiar/python-quick-start-clean-and-pickle-data\">script</a> to convert, clean and save the data as a pickle file. The preprocessing shouldn't throw away (or change) any usefull information.</p>\n\n<p>A quick summary:</p>\n\n<ul>\n<li>Convert json object to columns</li>\n<li>Drop constant columns (single unique value)</li>\n<li>Also drop campaignCode column (all missing except one row)</li>\n<li>Set datatype for a few numerical columns</li>\n<li>Save as pickle file</li>\n</ul>\n\n<p>To load the data: pandas.read_pickle('file_name'). If you find anything wrong please leave a comment.</p>",
      "rawMarkdown": "I did a [script][1] to convert, clean and save the data as a pickle file. The preprocessing shouldn't throw away (or change) any usefull information.\n\nA quick summary:\n\n - Convert json object to columns\n - Drop constant columns (single unique value)\n - Also drop campaignCode column (all missing except one row)\n - Set datatype for a few numerical columns\n - Save as pickle file\n\nTo load the data: pandas.read_pickle('file_name'). If you find anything wrong please leave a comment.\n\n\n  [1]: https://www.kaggle.com/jsaguiar/python-quick-start-clean-and-pickle-data",
      "votes": null
    },
    {
      "id": "388034",
      "postDate": "09/16/2018 06:31:32",
      "content": "<p>Why is campaign code missing in test data?</p>",
      "rawMarkdown": "Why is campaign code missing in test data?",
      "votes": null
    },
    {
      "id": "388237",
      "postDate": "09/16/2018 16:02:18",
      "content": "<p>I don't know why, but if you look at the values in train data they are all missing, except one single visit. This  unique visit is from the Philipines and has zero revenue, so I think it's totally safe to drop the column</p>",
      "rawMarkdown": "I don't know why, but if you look at the values in train data they are all missing, except one single visit. This  unique visit is from the Philipines and has zero revenue, so I think it's totally safe to drop the column",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 388034,
      "author_name": "suchith0312",
      "author_url": "",
      "post_date": "09/16/2018 06:31:32",
      "content": "<p>Why is campaign code missing in test data?</p>",
      "votes": null,
      "replies": [
        {
          "id": 388237,
          "author_name": "jsaguiar",
          "author_url": "",
          "post_date": "09/16/2018 16:02:18",
          "content": "<p>I don't know why, but if you look at the values in train data they are all missing, except one single visit. This  unique visit is from the Philipines and has zero revenue, so I think it's totally safe to drop the column</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "387983": "I did a [script][1] to convert, clean and save the data as a pickle file. The preprocessing shouldn't throw away (or change) any usefull information.\n\nA quick summary:\n\n - Convert json object to columns\n - Drop constant columns (single unique value)\n - Also drop campaignCode column (all missing except one row)\n - Set datatype for a few numerical columns\n - Save as pickle file\n\nTo load the data: pandas.read_pickle('file_name'). If you find anything wrong please leave a comment.\n\n\n  [1]: https://www.kaggle.com/jsaguiar/python-quick-start-clean-and-pickle-data",
    "388034": "Why is campaign code missing in test data?",
    "388237": "I don't know why, but if you look at the values in train data they are all missing, except one single visit. This  unique visit is from the Philipines and has zero revenue, so I think it's totally safe to drop the column"
  },
  "source": "meta"
}