{
  "fetched_at": "2026-10-19T10:33:20Z",
  "source_url": "https://www.kaggle.com/competitions/applications-of-deep-learning-wustlfall-2021/data",
  "pages": [
    {
      "name": "data-description",
      "content": "## Files\n\n*   **train.csv** - the training set\n*   **test.csv** - the test set\n\n## Columns\n\n*   `id` - The `id` of this city, for example, `id` 1 corresponds to the image in `1.jpg`.\n*   `sqft` - square feet of the city in millions of square feet.\n\n## Other notes\n\n* Remember your target is an area, not a volume.  The square feet, in millions, is equal to the sum of each building's `width * height * floors`.\n* The height of each floor is regulated by the nation to an undisclosed (yet uniform) number of feet.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "![Virtual City](https://data.heatonresearch.com/images/wustl/kaggle/city-size/orbit-107.jpg)\nThere are many different ways that engineers measure cities.  What is the population?  What is the total land area occupied by the city.  In this exercise, we will see if you can create a model to predict the square feet contained in various virtual cities.\n\nYou are provided with images looking towards the city center.  Each image is 512x512, compressed with JPEG.  The following sample image shows a typical view of a virtual city.  You can see towers of a variety of sizes and colors. \n\n![Typical City View](https://data.heatonresearch.com/images/wustl/kaggle/city-size/2.jpg)\n\nNot all views of the city will be the same. The following view looks at a city much closer up; however, results in some cropping of important detail.\n\n![Up Close](https://data.heatonresearch.com/images/wustl/kaggle/city-size/10.jpg)\n\nSome of the images will show you the city from somewhat above.  Here you can see the city roads more clearly.\n![Above the City](https://data.heatonresearch.com/images/wustl/kaggle/city-size/321.jpg)\n\nSome cities are very small.\n\n![Small City](https://data.heatonresearch.com/images/wustl/kaggle/city-size/350.jpg)\n\nOther cities are larger.\n\n![Large City](https://data.heatonresearch.com/images/wustl/kaggle/city-size/947.jpg)\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Kaggle will evaluate your submission with [RMSE](https://en.wikipedia.org/wiki/Root-mean-square_deviation).\n\n## Benchmark Scores\n\nThere are two benchmark scores I created in the Leaderboard.\n\n* **Average of Training Set** - This is the score you can get if you simply take the average all training data `sqft` and repeat that value for every test set row.\n* **Transfer Learning Starter** - This is the score that you can get by submitting the transfer learning example I placed in the code section.\n\n\n## Submission Format\n\nThe file should contain a header and have the following format:\n```\nid,sqft\n1,3240\n```",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "* Teams must be 5 people or less.\n* It is okay to compete by yourself.\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}