{
  "id": 243843,
  "title": "Is a 'map' a dataset?",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/243843",
  "author_name": "",
  "post_date": "2021-06-04T08:24:40.515107100Z",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Or more precisely: 'when is/isn't a 'map'  a dataset? <br>\nI would argue even a 'literal' map, if referred in a publication, refers to data that is used in the research thus should be considered as a 'dataset' . </p>",
  "messages": [
    {
      "id": "1335450",
      "postDate": "06/04/2021 08:24:40",
      "content": "<p>Or more precisely: 'when is/isn't a 'map'  a dataset? <br>\nI would argue even a 'literal' map, if referred in a publication, refers to data that is used in the research thus should be considered as a 'dataset' . </p>",
      "rawMarkdown": "Or more precisely: 'when is/isn't a 'map'  a dataset? \nI would argue even a 'literal' map, if referred in a publication, refers to data that is used in the research thus should be considered as a 'dataset' .",
      "votes": null
    },
    {
      "id": "1335770",
      "postDate": "06/04/2021 12:30:03",
      "content": "<p>Good question! I personally only define a map as a dataset in case the data behind are also accessible (on the webpage for instance). But that is just a subjective approach, there hasn't been any formal definition what constitutes a dataset by the host..</p>",
      "rawMarkdown": "Good question! I personally only define a map as a dataset in case the data behind are also accessible (on the webpage for instance). But that is just a subjective approach, there hasn't been any formal definition what constitutes a dataset by the host..",
      "votes": null
    },
    {
      "id": "1335796",
      "postDate": "06/04/2021 12:55:08",
      "content": "<p>… Good to see an experienced 'coloridger' around :-) <br>\nAnd I guess what you describe can't be really accomplished within ML approaches: How would the algorithm make this distinction, unless it has seen that specific map in its language model training by accident?<br>\n….. wait.. but that will still not do because the model will still probably not know any details in the pre-training. </p>",
      "rawMarkdown": "... Good to see an experienced 'coloridger' around :-) \nAnd I guess what you describe can't be really accomplished within ML approaches: How would the algorithm make this distinction, unless it has seen that specific map in its language model training by accident?\n..... wait.. but that will still not do because the model will still probably not know any details in the pre-training.",
      "votes": null
    },
    {
      "id": "1335991",
      "postDate": "06/04/2021 15:13:06",
      "content": "<p>You are right, this may only be useful for string matching approaches :-)</p>",
      "rawMarkdown": "You are right, this may only be useful for string matching approaches :-)",
      "votes": null
    },
    {
      "id": "1336229",
      "postDate": "06/04/2021 18:22:19",
      "content": "<p>I think that when maps like this \"usda plant hardiness zone map (phzm)\" are mentioned, they are talking about a dataset.</p>",
      "rawMarkdown": "I think that when maps like this \"usda plant hardiness zone map (phzm)\" are mentioned, they are talking about a dataset.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1335770,
      "author_name": "hannes82",
      "author_url": "",
      "post_date": "06/04/2021 12:30:03",
      "content": "<p>Good question! I personally only define a map as a dataset in case the data behind are also accessible (on the webpage for instance). But that is just a subjective approach, there hasn't been any formal definition what constitutes a dataset by the host..</p>",
      "votes": null,
      "replies": [
        {
          "id": 1335796,
          "author_name": "abdulkadirguner",
          "author_url": "",
          "post_date": "06/04/2021 12:55:08",
          "content": "<p>… Good to see an experienced 'coloridger' around :-) <br>\nAnd I guess what you describe can't be really accomplished within ML approaches: How would the algorithm make this distinction, unless it has seen that specific map in its language model training by accident?<br>\n….. wait.. but that will still not do because the model will still probably not know any details in the pre-training. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1335991,
          "author_name": "hannes82",
          "author_url": "",
          "post_date": "06/04/2021 15:13:06",
          "content": "<p>You are right, this may only be useful for string matching approaches :-)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1336229,
      "author_name": "mlconsult",
      "author_url": "",
      "post_date": "06/04/2021 18:22:19",
      "content": "<p>I think that when maps like this \"usda plant hardiness zone map (phzm)\" are mentioned, they are talking about a dataset.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1335450": "Or more precisely: 'when is/isn't a 'map'  a dataset? \nI would argue even a 'literal' map, if referred in a publication, refers to data that is used in the research thus should be considered as a 'dataset' .",
    "1335770": "Good question! I personally only define a map as a dataset in case the data behind are also accessible (on the webpage for instance). But that is just a subjective approach, there hasn't been any formal definition what constitutes a dataset by the host..",
    "1335796": "... Good to see an experienced 'coloridger' around :-) \nAnd I guess what you describe can't be really accomplished within ML approaches: How would the algorithm make this distinction, unless it has seen that specific map in its language model training by accident?\n..... wait.. but that will still not do because the model will still probably not know any details in the pre-training.",
    "1335991": "You are right, this may only be useful for string matching approaches :-)",
    "1336229": "I think that when maps like this \"usda plant hardiness zone map (phzm)\" are mentioned, they are talking about a dataset."
  },
  "source": "meta"
}