{
  "id": 248809,
  "title": "Anyone tried using Twitter?",
  "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/248809",
  "author_name": "",
  "post_date": "2021-06-25T04:49:07.312406500Z",
  "votes": 5,
  "comment_count": 2,
  "views": 0,
  "content": "<p>since we are trying to predict player's engagement. I thought Player's Twitter data gonna be an important feature so I decided to use it as a feature but so far I am not getting any significant improvement<br>\ndid anyone manage to use Player's Twitter Data effectively?</p>",
  "messages": [
    {
      "id": "1364554",
      "postDate": "06/25/2021 04:49:07",
      "content": "<p>since we are trying to predict player's engagement. I thought Player's Twitter data gonna be an important feature so I decided to use it as a feature but so far I am not getting any significant improvement<br>\ndid anyone manage to use Player's Twitter Data effectively?</p>",
      "rawMarkdown": "since we are trying to predict player's engagement. I thought Player's Twitter data gonna be an important feature so I decided to use it as a feature but so far I am not getting any significant improvement\ndid anyone manage to use Player's Twitter Data effectively?",
      "votes": null
    },
    {
      "id": "1365443",
      "postDate": "06/25/2021 18:08:52",
      "content": "<p>I did try but does not seem to help - I also did some analysis of twitter metrics and it does not seem to correlate with the targets.  I think the best way is using descriptive statistics of each players engagement history with the player box score data.</p>\n<p>Here is a notebook with this approach</p>\n<p><a href=\"https://www.kaggle.com/mlconsult/1-35-lightgbm-ann\" target=\"_blank\">https://www.kaggle.com/mlconsult/1-35-lightgbm-ann</a></p>\n<p>Here is the Twitter comparison engagement with targets</p>\n<p><a href=\"https://www.kaggle.com/mlconsult/twitter-data-compare\" target=\"_blank\">https://www.kaggle.com/mlconsult/twitter-data-compare</a></p>\n<p>Also see these discussions</p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499</a></p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474</a></p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111</a></p>",
      "rawMarkdown": "I did try but does not seem to help - I also did some analysis of twitter metrics and it does not seem to correlate with the targets.  I think the best way is using descriptive statistics of each players engagement history with the player box score data.\n\nHere is a notebook with this approach\n\nhttps://www.kaggle.com/mlconsult/1-35-lightgbm-ann\n\nHere is the Twitter comparison engagement with targets\n\nhttps://www.kaggle.com/mlconsult/twitter-data-compare\n\nAlso see these discussions\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111",
      "votes": null
    },
    {
      "id": "1368171",
      "postDate": "06/28/2021 11:27:29",
      "content": "<p><a href=\"https://www.kaggle.com/mlconsult\" target=\"_blank\">@mlconsult</a> Thanks for sharing those notebooks. </p>",
      "rawMarkdown": "mlconsult Thanks for sharing those notebooks.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1365443,
      "author_name": "mlconsult",
      "author_url": "",
      "post_date": "06/25/2021 18:08:52",
      "content": "<p>I did try but does not seem to help - I also did some analysis of twitter metrics and it does not seem to correlate with the targets.  I think the best way is using descriptive statistics of each players engagement history with the player box score data.</p>\n<p>Here is a notebook with this approach</p>\n<p><a href=\"https://www.kaggle.com/mlconsult/1-35-lightgbm-ann\" target=\"_blank\">https://www.kaggle.com/mlconsult/1-35-lightgbm-ann</a></p>\n<p>Here is the Twitter comparison engagement with targets</p>\n<p><a href=\"https://www.kaggle.com/mlconsult/twitter-data-compare\" target=\"_blank\">https://www.kaggle.com/mlconsult/twitter-data-compare</a></p>\n<p>Also see these discussions</p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499</a></p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474</a></p>\n<p><a href=\"https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111\" target=\"_blank\">https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 1368171,
          "author_name": "crained",
          "author_url": "",
          "post_date": "06/28/2021 11:27:29",
          "content": "<p><a href=\"https://www.kaggle.com/mlconsult\" target=\"_blank\">@mlconsult</a> Thanks for sharing those notebooks. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1364554": "since we are trying to predict player's engagement. I thought Player's Twitter data gonna be an important feature so I decided to use it as a feature but so far I am not getting any significant improvement\ndid anyone manage to use Player's Twitter Data effectively?",
    "1365443": "I did try but does not seem to help - I also did some analysis of twitter metrics and it does not seem to correlate with the targets.  I think the best way is using descriptive statistics of each players engagement history with the player box score data.\n\nHere is a notebook with this approach\n\nhttps://www.kaggle.com/mlconsult/1-35-lightgbm-ann\n\nHere is the Twitter comparison engagement with targets\n\nhttps://www.kaggle.com/mlconsult/twitter-data-compare\n\nAlso see these discussions\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/246499\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/248474\n\nhttps://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/discussion/247111",
    "1368171": "mlconsult Thanks for sharing those notebooks."
  },
  "source": "meta"
}