As a newbie, I was sad every time I read someone mention something along the lines that a pure tabular data competition hadn't been around for years.
\nI'm ready for another one!
\nWhere are you all going to next?
\nI doubt I'll do any image ones. I am seriously considering: https://www.kaggle.com/competitions/open-problems-multimodal/overview
\nPs: If a fun competition is found, I'm very interested in teaming up this time. :)
", "messages": [ { "id": 1915147, "postDate": "2022-08-26T17:51:59.550Z", "content": "I'll probably (im)patiently wait for the next tabular one, I have a hard time getting excited about non-tabular for some reason. Hope it isn't a long wait!
", "rawMarkdown": "I'll probably (im)patiently wait for the next tabular one, I have a hard time getting excited about non-tabular for some reason. Hope it isn't a long wait!", "votes": 5, "replies": [ { "id": 1915230, "postDate": "2022-08-26T19:25:38.690Z", "content": "Me too. I used to look at NLP competitions too but now it seems all backbone NLP models are getting very large. If I want to do any of them I will either have to upgrade my machine or subscribe to cloud service ☹️.
", "rawMarkdown": "Me too. I used to look at NLP competitions too but now it seems all backbone NLP models are getting very large. If I want to do any of them I will either have to upgrade my machine or subscribe to cloud service ☹️. ", "votes": 3 }, { "id": 1915256, "postDate": "2022-08-26T20:01:24.623Z", "content": "Yeah - those competitions are also somewhat interesting to me, but definitely feel less accessible. And at the end of the day I really like feature engineering and find it more rewarding than the the typical non-tabular/unstructured workflows.
", "rawMarkdown": "Yeah - those competitions are also somewhat interesting to me, but definitely feel less accessible. And at the end of the day I really like feature engineering and find it more rewarding than the the typical non-tabular/unstructured workflows." }, { "id": 1915307, "postDate": "2022-08-26T21:11:58.480Z", "content": "I suggest one could try this one- https://www.kaggle.com/competitions/dfl-bundesliga-data-shootout
\nIt seems like a different competition and is quite interesting in my opinion @aquatic
As a newbie, I was sad every time I read someone mention something along the lines that a pure tabular data competition hadn't been around for years.
\nI'm ready for another one!
\nWhere are you all going to next?
\nI doubt I'll do any image ones. I am seriously considering: https://www.kaggle.com/competitions/open-problems-multimodal/overview
\nPs: If a fun competition is found, I'm very interested in teaming up this time. :)
", "rawMarkdown": "As a newbie, I was sad every time I read someone mention something along the lines that a pure tabular data competition hadn't been around for years.\n\nI'm ready for another one!\n\nWhere are you all going to next?\n\nI doubt I'll do any image ones. I am seriously considering: https://www.kaggle.com/competitions/open-problems-multimodal/overview\n\nPs: If a fun competition is found, I'm very interested in teaming up this time. :)", "votes": 5 }, { "id": 1915129, "postDate": "2022-08-26T17:27:42.377Z", "content": "Dear @roberthatch
\nThe other day I noticed a Community competition being held by Fiverr; binary classification with small datasets that caught my eye. Note that it is a Community comp. so there are no medals or points etc. but personally that doesn't bother me at all…
\nIt closes in 20 days, so: short and sweet!
All the best,
\ncarl
I need one more medal though! Lol. (Doesn't feel right to have a silver but still need one more medal to move to expert :p)
", "rawMarkdown": "I need one more medal though! Lol. (Doesn't feel right to have a silver but still need one more medal to move to expert :p)", "votes": 2 }, { "id": 1915168, "postDate": "2022-08-26T18:06:49.540Z", "content": "The Fiverr challenge is good actually, small dataset and tabular one without many nulls. It is also an imbalanced data problem. I recommend it as well!
", "rawMarkdown": "The Fiverr challenge is good actually, small dataset and tabular one without many nulls. It is also an imbalanced data problem. I recommend it as well! ", "votes": 3 }, { "id": 1915493, "postDate": "2022-08-27T02:41:10.843Z", "content": "Fiverr is a nice one as it doesn't require large computing power to tests the lessons from the Amex competition. Notebooks running on kaggle is enough.
", "rawMarkdown": "Fiverr is a nice one as it doesn't require large computing power to tests the lessons from the Amex competition. Notebooks running on kaggle is enough. ", "votes": 1 } ] }, { "id": 1917662, "postDate": "2022-08-29T00:50:32.957Z", "content": "There are many fun ML/DL puzzles in the Capture the Flag comp here. Half are related to tabular data and half image or NLP.
", "rawMarkdown": "There are many fun ML/DL puzzles in the Capture the Flag comp [here][1]. Half are related to tabular data and half image or NLP.\n\n[1]: https://www.kaggle.com/competitions/ai-village-ctf", "votes": 1 }, { "id": 1915142, "postDate": "2022-08-26T17:46:18.523Z", "content": "I'm also looking at Open Problems - Multimodal Single-Cell Integration, mainly because @ambrosm has already posted one of their famous \"EDAs which make sense\", so you know there's something good to learn.
\nThe size of the data is scary (~100GB), but I already see a notebook explaining how reduce it significantly (will @sbunzini be our @raddar this time? :D).
But before starting a new ranked competition, this month's Tabular Playground Series is a nice palate cleanser: linear models, data fits in ram, CV takes 5 seconds to complete.
", "rawMarkdown": "I'm also looking at [Open Problems - Multimodal Single-Cell Integration](https://www.kaggle.com/competitions/open-problems-multimodal/overview), mainly because @ambrosm has already posted one of their famous \"EDAs which make sense\", so you know there's something good to learn.\nThe size of the data is scary (~100GB), but I already see [a notebook](https://www.kaggle.com/code/sbunzini/reduce-memory-usage-by-95-with-sparse-matrices/notebook) explaining how reduce it significantly (will @sbunzini be our @raddar this time? :D).\n\nBut before starting a new ranked competition, this month's Tabular Playground Series is a nice palate cleanser: linear models, data fits in ram, CV takes 5 seconds to complete.", "votes": 2, "replies": [ { "id": 1915149, "postDate": "2022-08-26T17:55:28.433Z", "content": "I admit I also might just stick with this one alone for another month or three. All the best ideas becoming public information is like Christmas - there's a lot of room for learning and optimizing regardless of it being \"closed\".
\nBut I should try my first Playground series, yes :)
", "rawMarkdown": "I admit I also might just stick with this one alone for another month or three. All the best ideas becoming public information is like Christmas - there's a lot of room for learning and optimizing regardless of it being \"closed\". \n\nBut I should try my first Playground series, yes :)", "votes": 1 }, { "id": 1915156, "postDate": "2022-08-26T18:00:38.900Z", "content": "From my experience from this month's playground series competition, I suggest you could perhaps skip and wait for 4-5 days for the next installment @roberthatch
", "rawMarkdown": "From my experience from this month's playground series competition, I suggest you could perhaps skip and wait for 4-5 days for the next installment @roberthatch ", "votes": 2 }, { "id": 1915163, "postDate": "2022-08-26T18:03:14.040Z", "content": "I agree, I think that at this point it has clearly become a noise (over)fitting competition. Making a simple high-bias low-variance model in a couple days has been fun, but I wouldn't dedicate more than a week to it.
", "rawMarkdown": "I agree, I think that at this point it has clearly become a noise (over)fitting competition. Making a simple high-bias low-variance model in a couple days has been fun, but I wouldn't dedicate more than a week to it.", "votes": 1 } ] }, { "id": 1975559, "postDate": "2022-10-06T20:54:38.507Z", "content": "In the end, the next one I'm doing is this protein one from Novozymes: https://www.kaggle.com/competitions/novozymes-enzyme-stability-prediction
\nI am also looking for teammate(s), though I might be a bit picky, I admit I have my sight set on Gold and $$$ in that one. ;)
", "rawMarkdown": "In the end, the next one I'm doing is this protein one from Novozymes: https://www.kaggle.com/competitions/novozymes-enzyme-stability-prediction\n\nI am also looking for teammate(s), though I might be a bit picky, I admit I have my sight set on Gold and $$$ in that one. ;)" }, { "id": 1921609, "postDate": "2022-09-01T00:43:30.263Z", "content": "Hi @roberthatch May I invite you to participate in this survey regarding your experience on Kaggle (10 min)? This is not a scam. We are a group of researchers at the City University of Hong Kong. The survey link is: https://cityuhk.questionpro.com/survey-of-kaggle-contestants
", "rawMarkdown": "Hi @roberthatch May I invite you to participate in this survey regarding your experience on Kaggle (10 min)? This is not a scam. We are a group of researchers at the City University of Hong Kong. The survey link is: https://cityuhk.questionpro.com/survey-of-kaggle-contestants" }, { "id": 1915430, "postDate": "2022-08-27T01:20:50.593Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 1915745, "postDate": "2022-08-27T09:34:03.893Z", "content": "Even I plan to revisit my approach, analyze the mistakes and try to redevelop the solution using top approaches. This I hope will help me improve my skills in this regard. In the meanwhile, I do have community competitions and playground editions to augment my knowledge too @yiwangnz !!
", "rawMarkdown": "Even I plan to revisit my approach, analyze the mistakes and try to redevelop the solution using top approaches. This I hope will help me improve my skills in this regard. In the meanwhile, I do have community competitions and playground editions to augment my knowledge too @yiwangnz !!", "votes": 1 } ] } ], "comments": [ { "id": 1915147, "author_name": "Joe Eddy", "author_url": "", "post_date": "2022-08-26T17:51:59.550000", "content": "I'll probably (im)patiently wait for the next tabular one, I have a hard time getting excited about non-tabular for some reason. Hope it isn't a long wait!
", "votes": 5, "replies": [ { "id": 1915230, "author_name": "Tonghui Li", "author_url": "", "post_date": "2022-08-26T19:25:38.690000", "content": "Me too. I used to look at NLP competitions too but now it seems all backbone NLP models are getting very large. If I want to do any of them I will either have to upgrade my machine or subscribe to cloud service ☹️.
", "votes": 3, "replies": [] }, { "id": 1915256, "author_name": "Joe Eddy", "author_url": "", "post_date": "2022-08-26T20:01:24.623000", "content": "Yeah - those competitions are also somewhat interesting to me, but definitely feel less accessible. And at the end of the day I really like feature engineering and find it more rewarding than the the typical non-tabular/unstructured workflows.
", "votes": 0, "replies": [] }, { "id": 1915307, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2022-08-26T21:11:58.480000", "content": "I suggest one could try this one- https://www.kaggle.com/competitions/dfl-bundesliga-data-shootout
\nIt seems like a different competition and is quite interesting in my opinion @aquatic
Dear @roberthatch
\nThe other day I noticed a Community competition being held by Fiverr; binary classification with small datasets that caught my eye. Note that it is a Community comp. so there are no medals or points etc. but personally that doesn't bother me at all…
\nIt closes in 20 days, so: short and sweet!
All the best,
\ncarl
I need one more medal though! Lol. (Doesn't feel right to have a silver but still need one more medal to move to expert :p)
", "votes": 2, "replies": [] }, { "id": 1915168, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2022-08-26T18:06:49.540000", "content": "The Fiverr challenge is good actually, small dataset and tabular one without many nulls. It is also an imbalanced data problem. I recommend it as well!
", "votes": 3, "replies": [] }, { "id": 1915493, "author_name": "Elias", "author_url": "", "post_date": "2022-08-27T02:41:10.843000", "content": "Fiverr is a nice one as it doesn't require large computing power to tests the lessons from the Amex competition. Notebooks running on kaggle is enough.
", "votes": 1, "replies": [] } ] }, { "id": 1917662, "author_name": "Chris Deotte", "author_url": "", "post_date": "2022-08-29T00:50:32.957000", "content": "There are many fun ML/DL puzzles in the Capture the Flag comp here. Half are related to tabular data and half image or NLP.
", "votes": 1, "replies": [] }, { "id": 1915142, "author_name": "Andrea Cognolato", "author_url": "", "post_date": "2022-08-26T17:46:18.523000", "content": "I'm also looking at Open Problems - Multimodal Single-Cell Integration, mainly because @ambrosm has already posted one of their famous \"EDAs which make sense\", so you know there's something good to learn.
\nThe size of the data is scary (~100GB), but I already see a notebook explaining how reduce it significantly (will @sbunzini be our @raddar this time? :D).
But before starting a new ranked competition, this month's Tabular Playground Series is a nice palate cleanser: linear models, data fits in ram, CV takes 5 seconds to complete.
", "votes": 2, "replies": [ { "id": 1915149, "author_name": "Robert Hatch", "author_url": "", "post_date": "2022-08-26T17:55:28.433000", "content": "I admit I also might just stick with this one alone for another month or three. All the best ideas becoming public information is like Christmas - there's a lot of room for learning and optimizing regardless of it being \"closed\".
\nBut I should try my first Playground series, yes :)
", "votes": 1, "replies": [] }, { "id": 1915156, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2022-08-26T18:00:38.900000", "content": "From my experience from this month's playground series competition, I suggest you could perhaps skip and wait for 4-5 days for the next installment @roberthatch
", "votes": 2, "replies": [] }, { "id": 1915163, "author_name": "Andrea Cognolato", "author_url": "", "post_date": "2022-08-26T18:03:14.040000", "content": "I agree, I think that at this point it has clearly become a noise (over)fitting competition. Making a simple high-bias low-variance model in a couple days has been fun, but I wouldn't dedicate more than a week to it.
", "votes": 1, "replies": [] } ] }, { "id": 1975559, "author_name": "Robert Hatch", "author_url": "", "post_date": "2022-10-06T20:54:38.507000", "content": "In the end, the next one I'm doing is this protein one from Novozymes: https://www.kaggle.com/competitions/novozymes-enzyme-stability-prediction
\nI am also looking for teammate(s), though I might be a bit picky, I admit I have my sight set on Gold and $$$ in that one. ;)
", "votes": 0, "replies": [] }, { "id": 1921609, "author_name": "Yang Liu", "author_url": "", "post_date": "2022-09-01T00:43:30.263000", "content": "Hi @roberthatch May I invite you to participate in this survey regarding your experience on Kaggle (10 min)? This is not a scam. We are a group of researchers at the City University of Hong Kong. The survey link is: https://cityuhk.questionpro.com/survey-of-kaggle-contestants
", "votes": 0, "replies": [] }, { "id": 1915430, "author_name": "", "author_url": "", "post_date": "2022-08-27T01:20:50.593000", "content": "", "votes": 0, "replies": [ { "id": 1915745, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2022-08-27T09:34:03.893000", "content": "Even I plan to revisit my approach, analyze the mistakes and try to redevelop the solution using top approaches. This I hope will help me improve my skills in this regard. In the meanwhile, I do have community competitions and playground editions to augment my knowledge too @yiwangnz !!
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "1915147": "I'll probably (im)patiently wait for the next tabular one, I have a hard time getting excited about non-tabular for some reason. Hope it isn't a long wait!", "1915005": "As a newbie, I was sad every time I read someone mention something along the lines that a pure tabular data competition hadn't been around for years.\n\nI'm ready for another one!\n\nWhere are you all going to next?\n\nI doubt I'll do any image ones. I am seriously considering: https://www.kaggle.com/competitions/open-problems-multimodal/overview\n\nPs: If a fun competition is found, I'm very interested in teaming up this time. :)", "1915129": "Dear @roberthatch \n\nThe other day I noticed a [Community competition being held by Fiverr](https://www.kaggle.com/competitions/predict-potential-spammers-on-fiverr); binary classification with small datasets that caught my eye. Note that it is a Community comp. so there are no medals or points *etc.* but personally that doesn't bother me at all...\nIt closes in 20 days, so: short and sweet!\n\nAll the best,\ncarl", "1917662": "There are many fun ML/DL puzzles in the Capture the Flag comp [here][1]. Half are related to tabular data and half image or NLP.\n\n[1]: https://www.kaggle.com/competitions/ai-village-ctf", "1915142": "I'm also looking at [Open Problems - Multimodal Single-Cell Integration](https://www.kaggle.com/competitions/open-problems-multimodal/overview), mainly because @ambrosm has already posted one of their famous \"EDAs which make sense\", so you know there's something good to learn.\nThe size of the data is scary (~100GB), but I already see [a notebook](https://www.kaggle.com/code/sbunzini/reduce-memory-usage-by-95-with-sparse-matrices/notebook) explaining how reduce it significantly (will @sbunzini be our @raddar this time? :D).\n\nBut before starting a new ranked competition, this month's Tabular Playground Series is a nice palate cleanser: linear models, data fits in ram, CV takes 5 seconds to complete.", "1975559": "In the end, the next one I'm doing is this protein one from Novozymes: https://www.kaggle.com/competitions/novozymes-enzyme-stability-prediction\n\nI am also looking for teammate(s), though I might be a bit picky, I admit I have my sight set on Gold and $$$ in that one. ;)", "1921609": "Hi @roberthatch May I invite you to participate in this survey regarding your experience on Kaggle (10 min)? This is not a scam. We are a group of researchers at the City University of Hong Kong. The survey link is: https://cityuhk.questionpro.com/survey-of-kaggle-contestants", "1915430": "" } }