In this competition, you will be segmenting and classifying different cloud formation types that are seen from satellite images. This work will help scientists improve climate models by understanding the atmospheric conditions that lead to specific cloud formations.
\n\nPlease refer to this kernel for a quickstart on run-length decoding.
\n\nHave fun and good luck!
", "messages": [ { "id": 600796, "postDate": "2019-08-16T15:08:34.013Z", "content": "In this competition, you will be segmenting and classifying different cloud formation types that are seen from satellite images. This work will help scientists improve climate models by understanding the atmospheric conditions that lead to specific cloud formations.
\n\nPlease refer to this kernel for a quickstart on run-length decoding.
\n\nHave fun and good luck!
", "rawMarkdown": "In this competition, you will be segmenting and classifying different cloud formation types that are seen from satellite images. This work will help scientists improve climate models by understanding the atmospheric conditions that lead to specific cloud formations.\n\nPlease refer to [this kernel](https://www.kaggle.com/inversion/run-length-decoding-and-flower-submission) for a quickstart on run-length decoding.\n\nHave fun and good luck!", "votes": 17 }, { "id": 601634, "postDate": "2019-08-17T23:26:34.187Z", "content": "Really excited to see more initiative concerning climate science in Machine Learning. Climate Change has become an increasingly important subject to tackle for engineers and scientists (as well as pretty much everyone in fact), and I deeply believe ML will have a big role to play in that. I have seen research groups like MILA offering internships and publishing projects concerning the subject, but it is even more encouraging to see Kaggle offering its competition platform for advancing climate research. In my case, I'd be happy to participate in more of those competitions.
", "rawMarkdown": "Really excited to see more initiative concerning climate science in Machine Learning. Climate Change has become an increasingly important subject to tackle for engineers and scientists (as well as pretty much everyone in fact), and I deeply believe ML will have a big role to play in that. I have seen research groups like MILA offering [internships](https://mila.quebec/en/2019/02/interns-for-the-ai-and-climate-change-project/) and [publishing projects](https://mila.quebec/en/2019/06/visualizing-the-impact-of-climate-change/) concerning the subject, but it is even more encouraging to see Kaggle offering its competition platform for advancing climate research. In my case, I'd be happy to participate in more of those competitions. ", "votes": 11, "replies": [ { "id": 602133, "postDate": "2019-08-18T16:17:20.757Z", "content": "I agree 100%.
\n\nInteresting background of this competition . . . I have a friend who is an atmospheric scientist. He was at a conference in Germany and saw Stephan's presentation on this topic. My friend emailed me saying he thought this would be a perfect fit for Kaggle, and Maggie Demkin and I reached out to Stephan immediately. We were very pleased that he was open to hosting a competition (it meant extra work for him and Hauke Schulz!)
\n\nWe are very much committed to finding and hosting as many of these types of competitions as possible. Data scientists have an amazing skillset, and we want to match them with problems that make a positive difference in the world.
\n\nThe recent Humpback Whale Identification Challenge was another recent competition that, in the words of the sponsors, \"was a game changer\" in allowing marine biologists to understand whale migration patterns.
\n\nIt's amazing to be part of a community that is willing to work towards solving problem that have so much potential for doing good in the world.
", "rawMarkdown": "I agree 100%. \n\nInteresting background of this competition . . . I have a friend who is an atmospheric scientist. He was at a conference in Germany and saw Stephan's presentation on this topic. My friend emailed me saying he thought this would be a perfect fit for Kaggle, and Maggie Demkin and I reached out to Stephan immediately. We were very pleased that he was open to hosting a competition (it meant extra work for him and Hauke Schulz!)\n\nWe are very much committed to finding and hosting as many of these types of competitions as possible. Data scientists have an amazing skillset, and we want to match them with problems that make a positive difference in the world.\n\nThe recent [Humpback Whale Identification Challenge](https://www.kaggle.com/c/humpback-whale-identification) was another recent competition that, in the words of the sponsors, \"was a game changer\" in allowing marine biologists to understand whale migration patterns.\n\nIt's amazing to be part of a community that is willing to work towards solving problem that have so much potential for doing good in the world.", "votes": 16 } ] }, { "id": 655581, "postDate": "2019-10-23T08:07:22.167Z", "content": "Thank you for the guide!
\n\nBTW, is it fine to build a solution based on someone published notebook, modify it and then submit the result?
", "rawMarkdown": "Thank you for the guide! \n\nBTW, is it fine to build a solution based on someone published notebook, modify it and then submit the result?", "votes": 1 }, { "id": 644447, "postDate": "2019-10-08T20:43:17.527Z", "content": "Thanks for the information and the star kicker.\nInitiative much appreciated.
\n\nregards,\nAdrian
", "rawMarkdown": "Thanks for the information and the star kicker.\nInitiative much appreciated.\n\nregards,\nAdrian", "votes": 1 }, { "id": 646194, "postDate": "2019-10-11T00:37:20.350Z", "content": "Thank you for this topic. I choose this one as my first kaggle competition since it is quite interesting. Good luck to my kaggle journey!
", "rawMarkdown": "Thank you for this topic. I choose this one as my first kaggle competition since it is quite interesting. Good luck to my kaggle journey!" }, { "id": 642492, "postDate": "2019-10-06T06:52:31.250Z", "content": "Hi, apologies if this is a stupid/lazy question but: are there instances in the data where more than 1 segment in an image has the same class? For example a single image with multiple Sugar regions?
\n\nThanks!\nZ.
", "rawMarkdown": "Hi, apologies if this is a stupid/lazy question but: are there instances in the data where more than 1 segment in an image has the same class? For example a single image with multiple Sugar regions?\n\nThanks!\nZ.", "replies": [ { "id": 642687, "postDate": "2019-10-06T13:40:46.360Z", "content": "https://www.kaggle.com/aleksandradeis/understanding-clouds-eda
\n\nHer EDA shows that a very large number of segments exist. Many are small and the result of noise (mostly along the black vertical stripe).
", "rawMarkdown": "https://www.kaggle.com/aleksandradeis/understanding-clouds-eda\n\nHer EDA shows that a very large number of segments exist. Many are small and the result of noise (mostly along the black vertical stripe)." } ] }, { "id": 663135, "postDate": "2019-11-01T13:48:31.817Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 600844, "postDate": "2019-08-16T16:21:27.947Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 601634, "author_name": "xhlulu", "author_url": "", "post_date": "2019-08-17T23:26:34.187000", "content": "Really excited to see more initiative concerning climate science in Machine Learning. Climate Change has become an increasingly important subject to tackle for engineers and scientists (as well as pretty much everyone in fact), and I deeply believe ML will have a big role to play in that. I have seen research groups like MILA offering internships and publishing projects concerning the subject, but it is even more encouraging to see Kaggle offering its competition platform for advancing climate research. In my case, I'd be happy to participate in more of those competitions.
", "votes": 11, "replies": [ { "id": 602133, "author_name": "inversion", "author_url": "", "post_date": "2019-08-18T16:17:20.757000", "content": "I agree 100%.
\n\nInteresting background of this competition . . . I have a friend who is an atmospheric scientist. He was at a conference in Germany and saw Stephan's presentation on this topic. My friend emailed me saying he thought this would be a perfect fit for Kaggle, and Maggie Demkin and I reached out to Stephan immediately. We were very pleased that he was open to hosting a competition (it meant extra work for him and Hauke Schulz!)
\n\nWe are very much committed to finding and hosting as many of these types of competitions as possible. Data scientists have an amazing skillset, and we want to match them with problems that make a positive difference in the world.
\n\nThe recent Humpback Whale Identification Challenge was another recent competition that, in the words of the sponsors, \"was a game changer\" in allowing marine biologists to understand whale migration patterns.
\n\nIt's amazing to be part of a community that is willing to work towards solving problem that have so much potential for doing good in the world.
", "votes": 16, "replies": [] } ] }, { "id": 655581, "author_name": "arctic", "author_url": "", "post_date": "2019-10-23T08:07:22.167000", "content": "Thank you for the guide!
\n\nBTW, is it fine to build a solution based on someone published notebook, modify it and then submit the result?
", "votes": 1, "replies": [] }, { "id": 644447, "author_name": "Adrian Zinovei", "author_url": "", "post_date": "2019-10-08T20:43:17.527000", "content": "Thanks for the information and the star kicker.\nInitiative much appreciated.
\n\nregards,\nAdrian
", "votes": 1, "replies": [] }, { "id": 646194, "author_name": "Jinghang", "author_url": "", "post_date": "2019-10-11T00:37:20.350000", "content": "Thank you for this topic. I choose this one as my first kaggle competition since it is quite interesting. Good luck to my kaggle journey!
", "votes": 0, "replies": [] }, { "id": 642492, "author_name": "Zylatis", "author_url": "", "post_date": "2019-10-06T06:52:31.250000", "content": "Hi, apologies if this is a stupid/lazy question but: are there instances in the data where more than 1 segment in an image has the same class? For example a single image with multiple Sugar regions?
\n\nThanks!\nZ.
", "votes": 0, "replies": [ { "id": 642687, "author_name": "PC Jimmmy", "author_url": "", "post_date": "2019-10-06T13:40:46.360000", "content": "https://www.kaggle.com/aleksandradeis/understanding-clouds-eda
\n\nHer EDA shows that a very large number of segments exist. Many are small and the result of noise (mostly along the black vertical stripe).
", "votes": 0, "replies": [] } ] }, { "id": 663135, "author_name": "", "author_url": "", "post_date": "2019-11-01T13:48:31.817000", "content": "", "votes": 1, "replies": [] }, { "id": 600844, "author_name": "", "author_url": "", "post_date": "2019-08-16T16:21:27.947000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "600796": "In this competition, you will be segmenting and classifying different cloud formation types that are seen from satellite images. This work will help scientists improve climate models by understanding the atmospheric conditions that lead to specific cloud formations.\n\nPlease refer to [this kernel](https://www.kaggle.com/inversion/run-length-decoding-and-flower-submission) for a quickstart on run-length decoding.\n\nHave fun and good luck!", "601634": "Really excited to see more initiative concerning climate science in Machine Learning. Climate Change has become an increasingly important subject to tackle for engineers and scientists (as well as pretty much everyone in fact), and I deeply believe ML will have a big role to play in that. I have seen research groups like MILA offering [internships](https://mila.quebec/en/2019/02/interns-for-the-ai-and-climate-change-project/) and [publishing projects](https://mila.quebec/en/2019/06/visualizing-the-impact-of-climate-change/) concerning the subject, but it is even more encouraging to see Kaggle offering its competition platform for advancing climate research. In my case, I'd be happy to participate in more of those competitions. ", "655581": "Thank you for the guide! \n\nBTW, is it fine to build a solution based on someone published notebook, modify it and then submit the result?", "644447": "Thanks for the information and the star kicker.\nInitiative much appreciated.\n\nregards,\nAdrian", "646194": "Thank you for this topic. I choose this one as my first kaggle competition since it is quite interesting. Good luck to my kaggle journey!", "642492": "Hi, apologies if this is a stupid/lazy question but: are there instances in the data where more than 1 segment in an image has the same class? For example a single image with multiple Sugar regions?\n\nThanks!\nZ.", "663135": "", "600844": "" } }