{
  "id": 613483,
  "title": "Start visualizing!",
  "url": "/competitions/physionet-ecg-image-digitization/discussion/613483",
  "author_name": "",
  "post_date": "2025-10-27T06:45:59.154733Z",
  "votes": 37,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I'm surprised that so many people develop, submit and even publish notebooks without visualizing the time series. Of today's 12 most popular public notebooks, the majority either contains no visualization at all or only the descriptive statistics we see in playground computations (e.g., a sample training image or a bar chart showing the frequency of image types).</p>\n<p>How can you practice computer vision without dissecting images? And how can you improve your time series predictions without plotting an overlay of y_true and y_pred?</p>\n<p>This competition needs visualizations of object detection:</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Ffe7df95bc1d5d81da51c144a16a140c9%2Fobject-detection.png?generation=1761547431542321&amp;alt=media\" alt=\"object detection\"></p>\n<p>And you want to see what parts of your predicted time series don't match the ground truth:</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Fa4186d02de58214c6d104ffaf6619996%2Ftrue-pred.png?generation=1761547443605641&amp;alt=media\" alt=\"true vs pred\"></p>\n<p>Source code is <a href=\"https://www.kaggle.com/code/ambrosm/ecg-original-explained-baseline\" target=\"_blank\">here</a>.</p>",
  "messages": [
    {
      "id": "3307506",
      "postDate": "10/27/2025 06:45:59",
      "content": "<p>I'm surprised that so many people develop, submit and even publish notebooks without visualizing the time series. Of today's 12 most popular public notebooks, the majority either contains no visualization at all or only the descriptive statistics we see in playground computations (e.g., a sample training image or a bar chart showing the frequency of image types).</p>\n<p>How can you practice computer vision without dissecting images? And how can you improve your time series predictions without plotting an overlay of y_true and y_pred?</p>\n<p>This competition needs visualizations of object detection:</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Ffe7df95bc1d5d81da51c144a16a140c9%2Fobject-detection.png?generation=1761547431542321&amp;alt=media\" alt=\"object detection\"></p>\n<p>And you want to see what parts of your predicted time series don't match the ground truth:</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Fa4186d02de58214c6d104ffaf6619996%2Ftrue-pred.png?generation=1761547443605641&amp;alt=media\" alt=\"true vs pred\"></p>\n<p>Source code is <a href=\"https://www.kaggle.com/code/ambrosm/ecg-original-explained-baseline\" target=\"_blank\">here</a>.</p>",
      "rawMarkdown": "I'm surprised that so many people develop, submit and even publish notebooks without visualizing the time series. Of today's 12 most popular public notebooks, the majority either contains no visualization at all or only the descriptive statistics we see in playground computations (e.g., a sample training image or a bar chart showing the frequency of image types).\n\nHow can you practice computer vision without dissecting images? And how can you improve your time series predictions without plotting an overlay of y_true and y_pred?\n\nThis competition needs visualizations of object detection:\n\n![object detection](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Ffe7df95bc1d5d81da51c144a16a140c9%2Fobject-detection.png?generation=1761547431542321&alt=media)\n\nAnd you want to see what parts of your predicted time series don't match the ground truth:\n\n![true vs pred](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Fa4186d02de58214c6d104ffaf6619996%2Ftrue-pred.png?generation=1761547443605641&alt=media)\n\nSource code is [here](https://www.kaggle.com/code/ambrosm/ecg-original-explained-baseline).",
      "votes": null
    },
    {
      "id": "3307529",
      "postDate": "10/27/2025 08:13:39",
      "content": "<p>I get it your means</p>",
      "rawMarkdown": "I get it your means",
      "votes": null
    },
    {
      "id": "3307754",
      "postDate": "10/27/2025 17:35:16",
      "content": "<p>Thank you for pointing that out. If you don’t mind, could you please share a general visualization notebook? I’m still working on improving my data analysis skills — especially for the current Playground series — and I’ve found very few notebooks that focus on visualization.</p>",
      "rawMarkdown": "Thank you for pointing that out. If you don’t mind, could you please share a general visualization notebook? I’m still working on improving my data analysis skills — especially for the current Playground series — and I’ve found very few notebooks that focus on visualization.",
      "votes": null
    },
    {
      "id": "3307934",
      "postDate": "10/28/2025 06:20:30",
      "content": "<p>Hi,</p>\n<p>Thanks for the point you've made. That makes a point, indeed.\nI'm just wondering, as a beginner, what are the meanings of those 17 endpoints you have implemented to find in the scanned image?</p>",
      "rawMarkdown": "Hi,\n\nThanks for the point you've made. That makes a point, indeed.\nI'm just wondering, as a beginner, what are the meanings of those 17 endpoints you have implemented to find in the scanned image?",
      "votes": null
    },
    {
      "id": "3307969",
      "postDate": "10/28/2025 08:58:29",
      "content": "<p>Thanks for this reminder, Ambros! It’s easy to get caught up in metrics and code, forgetting how much insight simple visualization brings.</p>",
      "rawMarkdown": "Thanks for this reminder, Ambros! It’s easy to get caught up in metrics and code, forgetting how much insight simple visualization brings.",
      "votes": null
    },
    {
      "id": "3308104",
      "postDate": "10/28/2025 15:31:49",
      "content": "<p>this the the truth!</p>",
      "rawMarkdown": "this the the truth!",
      "votes": null
    },
    {
      "id": "3321778",
      "postDate": "11/13/2025 06:53:38",
      "content": "<p>There are 13 different signals in the image (actually 12 since signal II appears twice.)  The 17 endpoints are just the start and end points of those 13 signal segments.</p>",
      "rawMarkdown": "There are 13 different signals in the image (actually 12 since signal II appears twice.)  The 17 endpoints are just the start and end points of those 13 signal segments.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3307529,
      "author_name": "qifeihhh666",
      "author_url": "",
      "post_date": "10/27/2025 08:13:39",
      "content": "<p>I get it your means</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3307754,
      "author_name": "gurseeon",
      "author_url": "",
      "post_date": "10/27/2025 17:35:16",
      "content": "<p>Thank you for pointing that out. If you don’t mind, could you please share a general visualization notebook? I’m still working on improving my data analysis skills — especially for the current Playground series — and I’ve found very few notebooks that focus on visualization.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3307934,
      "author_name": "deokjongmoon",
      "author_url": "",
      "post_date": "10/28/2025 06:20:30",
      "content": "<p>Hi,</p>\n<p>Thanks for the point you've made. That makes a point, indeed.\nI'm just wondering, as a beginner, what are the meanings of those 17 endpoints you have implemented to find in the scanned image?</p>",
      "votes": null,
      "replies": [
        {
          "id": 3321778,
          "author_name": "davidlist",
          "author_url": "",
          "post_date": "11/13/2025 06:53:38",
          "content": "<p>There are 13 different signals in the image (actually 12 since signal II appears twice.)  The 17 endpoints are just the start and end points of those 13 signal segments.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3307969,
      "author_name": "ladiposamson",
      "author_url": "",
      "post_date": "10/28/2025 08:58:29",
      "content": "<p>Thanks for this reminder, Ambros! It’s easy to get caught up in metrics and code, forgetting how much insight simple visualization brings.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3308104,
      "author_name": "irakozekelly",
      "author_url": "",
      "post_date": "10/28/2025 15:31:49",
      "content": "<p>this the the truth!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3307506": "I'm surprised that so many people develop, submit and even publish notebooks without visualizing the time series. Of today's 12 most popular public notebooks, the majority either contains no visualization at all or only the descriptive statistics we see in playground computations (e.g., a sample training image or a bar chart showing the frequency of image types).\n\nHow can you practice computer vision without dissecting images? And how can you improve your time series predictions without plotting an overlay of y_true and y_pred?\n\nThis competition needs visualizations of object detection:\n\n![object detection](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Ffe7df95bc1d5d81da51c144a16a140c9%2Fobject-detection.png?generation=1761547431542321&alt=media)\n\nAnd you want to see what parts of your predicted time series don't match the ground truth:\n\n![true vs pred](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7917824%2Fa4186d02de58214c6d104ffaf6619996%2Ftrue-pred.png?generation=1761547443605641&alt=media)\n\nSource code is [here](https://www.kaggle.com/code/ambrosm/ecg-original-explained-baseline).",
    "3307529": "I get it your means",
    "3307754": "Thank you for pointing that out. If you don’t mind, could you please share a general visualization notebook? I’m still working on improving my data analysis skills — especially for the current Playground series — and I’ve found very few notebooks that focus on visualization.",
    "3307934": "Hi,\n\nThanks for the point you've made. That makes a point, indeed.\nI'm just wondering, as a beginner, what are the meanings of those 17 endpoints you have implemented to find in the scanned image?",
    "3307969": "Thanks for this reminder, Ambros! It’s easy to get caught up in metrics and code, forgetting how much insight simple visualization brings.",
    "3308104": "this the the truth!",
    "3321778": "There are 13 different signals in the image (actually 12 since signal II appears twice.)  The 17 endpoints are just the start and end points of those 13 signal segments."
  },
  "source": "meta"
}