{
  "fetched_at": "2026-10-20T08:13:59Z",
  "source_url": "https://www.kaggle.com/competitions/cse465-can-you-predict-the-startup-industry/data",
  "pages": [
    {
      "name": "data-description",
      "content": "<p>Simple dataset collected from web scraping aims to provide an experimentation area for building startup industry classifier.</p>\n<p> Information provided in this dataset aims to be more general than participants think, there is urls provided in some columns that can point participant to more information through scrapping and other methods … keep your mind open for possible routes and solutions </p>\n  \n<h2>File descriptions</h2>\n\n- `train.csv` has all the data you will need for training and validation\n- `test.csv` what you need to predict\n- `sample.csv` a sample of how the submitted file should look like\n\n<h2>Data fields</h2>\n\n1. `name`: startup name\n2. `description`: startup self description \n3. `image`: url for the startup logo image\n4. `location`: startup headquarter location\n5. `stage`: startup current state (e.g. Idea, Generating Revenues and Scaling, etc.)\n6. `industry`: field of operation (target attribute that needs to be predicted in the test dataset)\n7. `subindustry`: more specific field of operation\n8. `website`: startup website",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<h2>Can we predict what is the field operation of a startup based on its description and meta-data ?</h2>\n\n<p>Sometimes startups can be vague in what they do and this derive most customers not to trust them or invest in their growth, do you think you can help with something like this ?</p>\n\nIn this competition you are tasked to build a **multi-class classifier** to predict the industry a startup is working on based on its available data on the internet.\n\n<h2>What should participants learn from this competition ? </h2>\n1. `sklearn` and its wide variety of algorithms and functionalities.\n2. `pandas` and how to preprocess data for better accuracy.\n3. Data Visualization using `matplotlib` or any other package.\n4. Working with small datasets and scrapping the internet for more information (`BeautifulSoup` or `scrapy`)\n5. Using kaggle notebooks for sharing code and ideas with others",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<p>As the goal of this competition is to predict the industries each startup is working on and that industries (by nature) are many, we found it suitable to use a MultiClass Evaluation Metric like <b>CategorizationAccuracy</b>.</p>\n\n<h2>Submission Format</h2>\nSubmission file should contain two columns: **id** and **industry**. The first column will be extracted from `test.csv` where the second column should be predicted from the classifier built (similar to `sample.csv`).\n \n<p>The file should contain a header and have the following format:</p>\n<pre>id,industry<br>1,Software and Data<br></pre>\n\n<h2>Some Tips</h2>\n1. Experiment with different encoding techniques.\n2. Analyse your data carefully before investing in any solution.\n3. Try as much classifiers as possible.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "This is a page where you can include rules that participants must accept before joining. You may wish to include rules like:\n\n - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}