{
  "id": 529412,
  "title": "Blackbody radiation of the stars",
  "url": "/competitions/ariel-data-challenge-2024/discussion/529412",
  "author_name": "Reza R. Choubeh",
  "post_date": "2024-08-20T16:22:23.314000",
  "votes": 4,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I plotted the intensity per wavelength of the two stars in the dataset before a transition. Since the detected wavelengths are in 2000-4000 nm range, I did not expect to recover the radiation spectrum of a balckbody including its peak, however, I expected that as the wavelengths get higher the intensity drops as is the case with a blackbody after its peak wavelength emission. <a href=\"https://www.e-education.psu.edu/meteo300/node/683\" target=\"_blank\">See here</a> for the spectra of Sun overlaid in an ideal blackbody spectrum.<br>\nThis is not the case. I am wondering why that is. One reason could be I applied the calibration the wrong way.<br>\nThe wavelengths in the plot below are in micro meter.<br>\nStar number zero:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2F140602def708add4c4d77d4bc3107375%2Fstar0.png?generation=1724170347346285&amp;alt=media\" alt=\"\"></p>\n<p>star number one:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2Feb8ee41117fc43d17586a34dc6d0a655%2Fstar1.png?generation=1724170385492694&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": 2965210,
      "postDate": "2024-08-20T16:22:23.313Z",
      "content": "<p>I plotted the intensity per wavelength of the two stars in the dataset before a transition. Since the detected wavelengths are in 2000-4000 nm range, I did not expect to recover the radiation spectrum of a balckbody including its peak, however, I expected that as the wavelengths get higher the intensity drops as is the case with a blackbody after its peak wavelength emission. <a href=\"https://www.e-education.psu.edu/meteo300/node/683\" target=\"_blank\">See here</a> for the spectra of Sun overlaid in an ideal blackbody spectrum.<br>\nThis is not the case. I am wondering why that is. One reason could be I applied the calibration the wrong way.<br>\nThe wavelengths in the plot below are in micro meter.<br>\nStar number zero:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2F140602def708add4c4d77d4bc3107375%2Fstar0.png?generation=1724170347346285&amp;alt=media\" alt=\"\"></p>\n<p>star number one:<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2Feb8ee41117fc43d17586a34dc6d0a655%2Fstar1.png?generation=1724170385492694&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "I plotted the intensity per wavelength of the two stars in the dataset before a transition. Since the detected wavelengths are in 2000-4000 nm range, I did not expect to recover the radiation spectrum of a balckbody including its peak, however, I expected that as the wavelengths get higher the intensity drops as is the case with a blackbody after its peak wavelength emission. [See here](https://www.e-education.psu.edu/meteo300/node/683) for the spectra of Sun overlaid in an ideal blackbody spectrum.\nThis is not the case. I am wondering why that is. One reason could be I applied the calibration the wrong way.\nThe wavelengths in the plot below are in micro meter.\nStar number zero:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2F140602def708add4c4d77d4bc3107375%2Fstar0.png?generation=1724170347346285&alt=media)\n\nstar number one:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2Feb8ee41117fc43d17586a34dc6d0a655%2Fstar1.png?generation=1724170385492694&alt=media)",
      "votes": 3
    },
    {
      "id": 2965289,
      "postDate": "2024-08-20T17:59:05.523Z",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/rezanl\" target=\"_blank\">@rezanl</a>, you have to flip left and right in the diagram. It's documented in the metadata that they installed the spectrometer in the satellite upside-down:</p>\n<pre><code> = pd.read_csv('/kaggle/input/ariel-data-challenge-/wavelengths.csv')\n = pd.read_parquet('/kaggle/input/ariel-data-challenge-/axis_info.parquet')\n(wavelength.values[,:])\n\n(axis_info['AIRS-CH0-axis2-um'].values[:])\n\n.scatter(wavelength.values[,:], a_train.mean(axis=)[][::-], s=)\n</code></pre>",
      "rawMarkdown": "Hi @rezanl, you have to flip left and right in the diagram. It's documented in the metadata that they installed the spectrometer in the satellite upside-down:\n```\nwavelength = pd.read_csv('/kaggle/input/ariel-data-challenge-2024/wavelengths.csv')\naxis_info = pd.read_parquet('/kaggle/input/ariel-data-challenge-2024/axis_info.parquet')\nprint(wavelength.values[0,1:])\n# [1.95 1.96 1.97 1.98 1.99 ... 3.88 3.88 3.89 3.89 3.9 ]\nprint(axis_info['AIRS-CH0-axis2-um'].values[39:321])\n# [3.9  3.89 3.89 3.88 3.88 ... 1.99 1.98 1.97 1.96 1.95]\nplt.scatter(wavelength.values[0,1:], a_train.mean(axis=1)[0][320:38:-1], s=5)\n```\n",
      "votes": 4,
      "replies": [
        {
          "id": 2965316,
          "postDate": "2024-08-20T18:34:56.903Z",
          "content": "<p>Thanks! You are right! </p>",
          "rawMarkdown": "Thanks! You are right! ",
          "votes": 1
        },
        {
          "id": 2978285,
          "postDate": "2024-09-03T17:51:44.577Z",
          "content": "<p>Hello, could you please explain what is a_train and why are you doing <code>a_train.mean(axis=1)[0][320:38:-1]</code>?</p>",
          "rawMarkdown": "Hello, could you please explain what is a_train and why are you doing `a_train.mean(axis=1)[0][320:38:-1]`?",
          "replies": [
            {
              "id": 2978349,
              "postDate": "2024-09-03T19:04:55.613Z",
              "content": "<p>Hi <a href=\"https://www.kaggle.com/markoster\" target=\"_blank\">@markoster</a> a_train is my name for the array of shape (n_planets, n_timesteps, n_wavelengths) containing the AIRS-CH0 measurements. <code>a_train.mean(axis=1)[0][320:38:-1]</code> averages over all timesteps, selects only the first planet, selects the subset of wavelengths which is relevant for the competition and finally flips left and right.</p>",
              "rawMarkdown": "Hi @markoster a_train is my name for the array of shape (n_planets, n_timesteps, n_wavelengths) containing the AIRS-CH0 measurements. `a_train.mean(axis=1)[0][320:38:-1]` averages over all timesteps, selects only the first planet, selects the subset of wavelengths which is relevant for the competition and finally flips left and right.",
              "votes": 1
            }
          ]
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2965289,
      "author_name": "AmbrosM",
      "author_url": "",
      "post_date": "2024-08-20T17:59:05.523000",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/rezanl\" target=\"_blank\">@rezanl</a>, you have to flip left and right in the diagram. It's documented in the metadata that they installed the spectrometer in the satellite upside-down:</p>\n<pre><code> = pd.read_csv('/kaggle/input/ariel-data-challenge-/wavelengths.csv')\n = pd.read_parquet('/kaggle/input/ariel-data-challenge-/axis_info.parquet')\n(wavelength.values[,:])\n\n(axis_info['AIRS-CH0-axis2-um'].values[:])\n\n.scatter(wavelength.values[,:], a_train.mean(axis=)[][::-], s=)\n</code></pre>",
      "votes": 4,
      "replies": [
        {
          "id": 2965316,
          "author_name": "Reza R. Choubeh",
          "author_url": "",
          "post_date": "2024-08-20T18:34:56.903000",
          "content": "<p>Thanks! You are right! </p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 2978285,
          "author_name": "Marcos Martinez",
          "author_url": "",
          "post_date": "2024-09-03T17:51:44.577000",
          "content": "<p>Hello, could you please explain what is a_train and why are you doing <code>a_train.mean(axis=1)[0][320:38:-1]</code>?</p>",
          "votes": 0,
          "replies": [
            {
              "id": 2978349,
              "author_name": "AmbrosM",
              "author_url": "",
              "post_date": "2024-09-03T19:04:55.613000",
              "content": "<p>Hi <a href=\"https://www.kaggle.com/markoster\" target=\"_blank\">@markoster</a> a_train is my name for the array of shape (n_planets, n_timesteps, n_wavelengths) containing the AIRS-CH0 measurements. <code>a_train.mean(axis=1)[0][320:38:-1]</code> averages over all timesteps, selects only the first planet, selects the subset of wavelengths which is relevant for the competition and finally flips left and right.</p>",
              "votes": 1,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2965210": "I plotted the intensity per wavelength of the two stars in the dataset before a transition. Since the detected wavelengths are in 2000-4000 nm range, I did not expect to recover the radiation spectrum of a balckbody including its peak, however, I expected that as the wavelengths get higher the intensity drops as is the case with a blackbody after its peak wavelength emission. [See here](https://www.e-education.psu.edu/meteo300/node/683) for the spectra of Sun overlaid in an ideal blackbody spectrum.\nThis is not the case. I am wondering why that is. One reason could be I applied the calibration the wrong way.\nThe wavelengths in the plot below are in micro meter.\nStar number zero:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2F140602def708add4c4d77d4bc3107375%2Fstar0.png?generation=1724170347346285&alt=media)\n\nstar number one:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6524939%2Feb8ee41117fc43d17586a34dc6d0a655%2Fstar1.png?generation=1724170385492694&alt=media)",
    "2965289": "Hi @rezanl, you have to flip left and right in the diagram. It's documented in the metadata that they installed the spectrometer in the satellite upside-down:\n```\nwavelength = pd.read_csv('/kaggle/input/ariel-data-challenge-2024/wavelengths.csv')\naxis_info = pd.read_parquet('/kaggle/input/ariel-data-challenge-2024/axis_info.parquet')\nprint(wavelength.values[0,1:])\n# [1.95 1.96 1.97 1.98 1.99 ... 3.88 3.88 3.89 3.89 3.9 ]\nprint(axis_info['AIRS-CH0-axis2-um'].values[39:321])\n# [3.9  3.89 3.89 3.88 3.88 ... 1.99 1.98 1.97 1.96 1.95]\nplt.scatter(wavelength.values[0,1:], a_train.mean(axis=1)[0][320:38:-1], s=5)\n```\n"
  }
}