Hi,
\n\nI've volunteered to speak at a local meetup, where we typically have a speaker present on a completed competition, and discuss the competition, and the methods which the winners used.
\n\nThe group is great, and presenting on completed Kaggle competitions produces great discussion and insight into what worked, or didn't work. Every second week, more than 50 people show up, usually with 1/4 to 1/2 being there for their first time. The number of people showing up is great, but we have a retention rate much lower than I would like.
\n\nIn order to get more people involved, I would like to try presenting on the thought processes and steps involved in entering a competition. It would effectively mimic a Kernel, except with the author of the Kernel walking everyone through it. Doing this on a live competition would have the added bonus of providing a room full of data science enthusiasts with the background info and starter kit of code to enter the competition themselves. Most attendants of the meetup have never competed themselves, and I believe this format would reduce the barrier to entry for anyone competing for the first time.
\n\nAll of my code would be published as a Kernel prior to the talk, the talk would be live streamed via YouTube, with a link to it posted here, and anyone in the live stream can ask questions during the presentation as well. Per the competition rules, \"Privately sharing code or data outside of teams is not permitted. It's okay to share code if made available to all participants on the forums.\" I am aware that this can be interpreted such that what I am proposing is not allowed. I am not trying to take advantage of any loophole in the rules, and understand if this is not allowed.
\n\nThe presentation will be on May 3rd, at 6:30pm. Videos of past competition presentations available here https://www.youtube.com/user/Gauss256
", "messages": [ { "id": "315483", "postDate": "04/17/2018 06:54:34", "content": "Hi,
\n\nI've volunteered to speak at a local meetup, where we typically have a speaker present on a completed competition, and discuss the competition, and the methods which the winners used.
\n\nThe group is great, and presenting on completed Kaggle competitions produces great discussion and insight into what worked, or didn't work. Every second week, more than 50 people show up, usually with 1/4 to 1/2 being there for their first time. The number of people showing up is great, but we have a retention rate much lower than I would like.
\n\nIn order to get more people involved, I would like to try presenting on the thought processes and steps involved in entering a competition. It would effectively mimic a Kernel, except with the author of the Kernel walking everyone through it. Doing this on a live competition would have the added bonus of providing a room full of data science enthusiasts with the background info and starter kit of code to enter the competition themselves. Most attendants of the meetup have never competed themselves, and I believe this format would reduce the barrier to entry for anyone competing for the first time.
\n\nAll of my code would be published as a Kernel prior to the talk, the talk would be live streamed via YouTube, with a link to it posted here, and anyone in the live stream can ask questions during the presentation as well. Per the competition rules, \"Privately sharing code or data outside of teams is not permitted. It's okay to share code if made available to all participants on the forums.\" I am aware that this can be interpreted such that what I am proposing is not allowed. I am not trying to take advantage of any loophole in the rules, and understand if this is not allowed.
\n\nThe presentation will be on May 3rd, at 6:30pm. Videos of past competition presentations available here https://www.youtube.com/user/Gauss256
", "rawMarkdown": "Hi,\n\nI've volunteered to speak at a local meetup, where we typically have a speaker present on a completed competition, and discuss the competition, and the methods which the winners used.\n\nThe group is great, and presenting on completed Kaggle competitions produces great discussion and insight into what worked, or didn't work. Every second week, more than 50 people show up, usually with 1/4 to 1/2 being there for their first time. The number of people showing up is great, but we have a retention rate much lower than I would like. \n\nIn order to get more people involved, I would like to try presenting on the thought processes and steps involved in entering a competition. It would effectively mimic a Kernel, except with the author of the Kernel walking everyone through it. Doing this on a live competition would have the added bonus of providing a room full of data science enthusiasts with the background info and starter kit of code to enter the competition themselves. Most attendants of the meetup have never competed themselves, and I believe this format would reduce the barrier to entry for anyone competing for the first time.\n\nAll of my code would be published as a Kernel prior to the talk, the talk would be live streamed via YouTube, with a link to it posted here, and anyone in the live stream can ask questions during the presentation as well. Per the competition rules, \"Privately sharing code or data outside of teams is not permitted. It's okay to share code if made available to all participants on the forums.\" I am aware that this can be interpreted such that what I am proposing is not allowed. I am not trying to take advantage of any loophole in the rules, and understand if this is not allowed.\n\nThe presentation will be on May 3rd, at 6:30pm. Videos of past competition presentations available here https://www.youtube.com/user/Gauss256", "votes": null }, { "id": "315920", "postDate": "04/17/2018 20:01:01", "content": "Hi Matt,
\n\nThanks for sharing your knowledge in these meetups! We love hearing our community organize events like this one. You're correct on the rule, however as you note that the Kernel is published publicly, the talk is live streamed (and I assume available for re-watch later), I do not see this as any different than hosting a \"Kaggle discussion forum\" in person instead of on the competition page, yes?
", "rawMarkdown": "Hi Matt,\n\nThanks for sharing your knowledge in these meetups! We love hearing our community organize events like this one. You're correct on the rule, however as you note that the Kernel is published publicly, the talk is live streamed (and I assume available for re-watch later), I do not see this as any different than hosting a \"Kaggle discussion forum\" in person instead of on the competition page, yes?", "votes": null }, { "id": "316030", "postDate": "04/18/2018 03:27:46", "content": "Yes, the video will be available for re-watch at any time.\nI was mostly just confirming that hosting in person was allowed, given that all information will be publicly available.
\n\nThanks!
", "rawMarkdown": "Yes, the video will be available for re-watch at any time.\nI was mostly just confirming that hosting in person was allowed, given that all information will be publicly available.\n\nThanks!", "votes": null }, { "id": "322897", "postDate": "05/03/2018 21:52:12", "content": "Just bumping this thread.
\n\nI've uploaded the notebook I will be following for my presentation this evening. It's just running right now, but should finish soon. Link to the Kernel here https://www.kaggle.com/mpotma/learndatascience-presentation-lgbm-lb-0-836
\n\nThe talk will begin around 6:30pm PDT this evening, more information about it here https://www.meetup.com/LearnDataScience/events/bmcntlyxhbfb/
\n\nI'll make sure that a link to the livestream is posted somewhere here shortly before the talk begins, so that anyone who wants to watch it live can do so. It'll also be saved for viewing at any time in the future.
", "rawMarkdown": "Just bumping this thread.\n\nI've uploaded the notebook I will be following for my presentation this evening. It's just running right now, but should finish soon. Link to the Kernel here https://www.kaggle.com/mpotma/learndatascience-presentation-lgbm-lb-0-836\n\nThe talk will begin around 6:30pm PDT this evening, more information about it here https://www.meetup.com/LearnDataScience/events/bmcntlyxhbfb/\n\nI'll make sure that a link to the livestream is posted somewhere here shortly before the talk begins, so that anyone who wants to watch it live can do so. It'll also be saved for viewing at any time in the future.", "votes": null }, { "id": "322940", "postDate": "05/04/2018 01:24:44", "content": "Unfortunately, my Kernel is still running, but the livestream link is right here\nhttps://www.youtube.com/channel/UCJhW_16uxALr0X4olEW2p5A/live
", "rawMarkdown": "Unfortunately, my Kernel is still running, but the livestream link is right here\nhttps://www.youtube.com/channel/UCJhW_16uxALr0X4olEW2p5A/live", "votes": null }, { "id": "323662", "postDate": "05/05/2018 20:05:05", "content": "Thanks for the interest in the competition! Organizing events like this to discuss Kaggle competitions seems like a great initiative.
", "rawMarkdown": "Thanks for the interest in the competition! Organizing events like this to discuss Kaggle competitions seems like a great initiative.", "votes": null } ], "comments": [ { "id": 315920, "author_name": "addisonhoward", "author_url": "", "post_date": "04/17/2018 20:01:01", "content": "Hi Matt,
\n\nThanks for sharing your knowledge in these meetups! We love hearing our community organize events like this one. You're correct on the rule, however as you note that the Kernel is published publicly, the talk is live streamed (and I assume available for re-watch later), I do not see this as any different than hosting a \"Kaggle discussion forum\" in person instead of on the competition page, yes?
", "votes": null, "replies": [ { "id": 316030, "author_name": "mpotma", "author_url": "", "post_date": "04/18/2018 03:27:46", "content": "Yes, the video will be available for re-watch at any time.\nI was mostly just confirming that hosting in person was allowed, given that all information will be publicly available.
\n\nThanks!
", "votes": null, "replies": [] } ] }, { "id": 322897, "author_name": "mpotma", "author_url": "", "post_date": "05/03/2018 21:52:12", "content": "Just bumping this thread.
\n\nI've uploaded the notebook I will be following for my presentation this evening. It's just running right now, but should finish soon. Link to the Kernel here https://www.kaggle.com/mpotma/learndatascience-presentation-lgbm-lb-0-836
\n\nThe talk will begin around 6:30pm PDT this evening, more information about it here https://www.meetup.com/LearnDataScience/events/bmcntlyxhbfb/
\n\nI'll make sure that a link to the livestream is posted somewhere here shortly before the talk begins, so that anyone who wants to watch it live can do so. It'll also be saved for viewing at any time in the future.
", "votes": null, "replies": [ { "id": 323662, "author_name": "eduardofonseca", "author_url": "", "post_date": "05/05/2018 20:05:05", "content": "Thanks for the interest in the competition! Organizing events like this to discuss Kaggle competitions seems like a great initiative.
", "votes": null, "replies": [] } ] }, { "id": 322940, "author_name": "mpotma", "author_url": "", "post_date": "05/04/2018 01:24:44", "content": "Unfortunately, my Kernel is still running, but the livestream link is right here\nhttps://www.youtube.com/channel/UCJhW_16uxALr0X4olEW2p5A/live
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "315483": "Hi,\n\nI've volunteered to speak at a local meetup, where we typically have a speaker present on a completed competition, and discuss the competition, and the methods which the winners used.\n\nThe group is great, and presenting on completed Kaggle competitions produces great discussion and insight into what worked, or didn't work. Every second week, more than 50 people show up, usually with 1/4 to 1/2 being there for their first time. The number of people showing up is great, but we have a retention rate much lower than I would like. \n\nIn order to get more people involved, I would like to try presenting on the thought processes and steps involved in entering a competition. It would effectively mimic a Kernel, except with the author of the Kernel walking everyone through it. Doing this on a live competition would have the added bonus of providing a room full of data science enthusiasts with the background info and starter kit of code to enter the competition themselves. Most attendants of the meetup have never competed themselves, and I believe this format would reduce the barrier to entry for anyone competing for the first time.\n\nAll of my code would be published as a Kernel prior to the talk, the talk would be live streamed via YouTube, with a link to it posted here, and anyone in the live stream can ask questions during the presentation as well. Per the competition rules, \"Privately sharing code or data outside of teams is not permitted. It's okay to share code if made available to all participants on the forums.\" I am aware that this can be interpreted such that what I am proposing is not allowed. I am not trying to take advantage of any loophole in the rules, and understand if this is not allowed.\n\nThe presentation will be on May 3rd, at 6:30pm. Videos of past competition presentations available here https://www.youtube.com/user/Gauss256", "315920": "Hi Matt,\n\nThanks for sharing your knowledge in these meetups! We love hearing our community organize events like this one. You're correct on the rule, however as you note that the Kernel is published publicly, the talk is live streamed (and I assume available for re-watch later), I do not see this as any different than hosting a \"Kaggle discussion forum\" in person instead of on the competition page, yes?", "316030": "Yes, the video will be available for re-watch at any time.\nI was mostly just confirming that hosting in person was allowed, given that all information will be publicly available.\n\nThanks!", "322897": "Just bumping this thread.\n\nI've uploaded the notebook I will be following for my presentation this evening. It's just running right now, but should finish soon. Link to the Kernel here https://www.kaggle.com/mpotma/learndatascience-presentation-lgbm-lb-0-836\n\nThe talk will begin around 6:30pm PDT this evening, more information about it here https://www.meetup.com/LearnDataScience/events/bmcntlyxhbfb/\n\nI'll make sure that a link to the livestream is posted somewhere here shortly before the talk begins, so that anyone who wants to watch it live can do so. It'll also be saved for viewing at any time in the future.", "322940": "Unfortunately, my Kernel is still running, but the livestream link is right here\nhttps://www.youtube.com/channel/UCJhW_16uxALr0X4olEW2p5A/live", "323662": "Thanks for the interest in the competition! Organizing events like this to discuss Kaggle competitions seems like a great initiative." }, "source": "meta" }