![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F19099%2F735ede9d63486dd72e041d7e1bb3009e%2F2022-11-10%20%2001.45.54.png?generation=1668033988689057&alt=media\")
Hi fellow Kagglers,
\nI have created the dataset and the example kernel to make some analytics on the kernels, which were created for this competition.
\nIn the kernel you can see several interesting points like:
\nTo make it more interesting to check here are some results.
\nHere we can congrats @cdeotte / @edwardcrookenden / @vslaykovsky / @simamumu who are dividing the current 1st place with 1 gold medal, @radek1 for the 2nd place with 4 silver medals and @ttahara for the 3rd place respectively with 2 silvers:
\n
The full kernels graph for this competition looks like:
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F19099%2Fb9f5d55b49eb9fa1c0b9af54e6ea173c%2F2022-11-10%20%2001.49.37.png?generation=1668034205080856&alt=media\")
I have used the kernel IDs in the graph to make it more readable so to find your kernel ID you can check it in the dataset by your name or kernel title.
\nIf we use this graph and calculate the cumulative number of upvotes using the forks, we can receive the table like:
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F19099%2F9ef8c82ff7d3b170a6c02833abb776ad%2F2022-11-10%20%2001.50.14.png?generation=1668034246721726&alt=media\")
The current winner is @simamumu with his [OLD] test_data last 20 aid get LB0.947 kernel. @vslaykovsky is on the 2nd place and the 3rd is @cdeotte with amazing found - test data leak, which can help you to boost the LB score.
The graph just says \"currently it's only the beginning of the competition - the score doesn't role so much now\" 🙃
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F19099%2Fb981999ad43992e1a9eff50e2f011e85%2F2022-11-10%20%2001.54.07.png?generation=1668034477435477&alt=media\")
Looking forward for your comments, suggestions and insights!
\nAlex
", "messages": [ { "id": "2023638", "postDate": "11/09/2022 22:54:52", "content": "Hi fellow Kagglers,
\nI have created the dataset and the example kernel to make some analytics on the kernels, which were created for this competition.
\nIn the kernel you can see several interesting points like:
\nTo make it more interesting to check here are some results.
\nHere we can congrats @cdeotte / @edwardcrookenden / @vslaykovsky / @simamumu who are dividing the current 1st place with 1 gold medal, @radek1 for the 2nd place with 4 silver medals and @ttahara for the 3rd place respectively with 2 silvers:
\n
The full kernels graph for this competition looks like:
\n
I have used the kernel IDs in the graph to make it more readable so to find your kernel ID you can check it in the dataset by your name or kernel title.
\nIf we use this graph and calculate the cumulative number of upvotes using the forks, we can receive the table like:
\n
The current winner is @simamumu with his [OLD] test_data last 20 aid get LB0.947 kernel. @vslaykovsky is on the 2nd place and the 3rd is @cdeotte with amazing found - test data leak, which can help you to boost the LB score.
The graph just says \"currently it's only the beginning of the competition - the score doesn't role so much now\" 🙃
\n
Looking forward for your comments, suggestions and insights!
\nAlex
", "rawMarkdown": "Hi fellow Kagglers,\n\nI have created [the dataset](https://www.kaggle.com/datasets/alexryzhkov/otto-recommender-kernel-statistics) and [the example kernel](https://www.kaggle.com/code/alexryzhkov/otto-recommender-kernel-stats) to make some analytics on the kernels, which were created for this competition.\n\nIn the kernel you can see several interesting points like:\n- Top-20 kernel creators\n- Performance Tiers of kernel creators\n- Kernels medals stats\n- Medals for kernels creators (top-20)\n- Kernels languages\n- How many kernels are the forks?\n- How many lines are changed in forks?\n- Does the kernel use GPU or not?\n- Upvotes vs views and Upvotes vs PublicScore\n- Most commented kernels\n- **Kernel forks graph** and tops on it\n\nTo make it more interesting to check here are some results.\n\n### Medals for kernels\n\nHere we can congrats @cdeotte / @edwardcrookenden / @vslaykovsky / @simamumu who are dividing the current 1st place with 1 gold medal, @radek1 for the 2nd place with 4 silver medals and @ttahara for the 3rd place respectively with 2 silvers:\n\n\n\n### Forks graph\n\nThe full kernels graph for this competition looks like:\n\n\n\nI have used the kernel IDs in the graph to make it more readable so to find your kernel ID you can check it in the dataset by your name or kernel title.\n\nIf we use this graph and calculate the cumulative number of upvotes using the forks, we can receive the table like:\n\n\n\nThe current winner is @simamumu with his [[OLD] test_data last 20 `aid` get LB0.947](https://www.kaggle.com/code/simamumu/old-test-data-last-20-aid-get-lb0-947) kernel. @vslaykovsky is on the 2nd place and the 3rd is @cdeotte with amazing found - test data leak, which can help you to boost the LB score.\n\n### Kernel votes vs public score\n\nThe graph just says \"currently it's only the beginning of the competition - the score doesn't role so much now\" 🙃\n\n\n\nLooking forward for your comments, suggestions and insights!\n\nAlex", "votes": null }, { "id": "2023724", "postDate": "11/10/2022 01:32:49", "content": "This is awesome stuff, wow! Really interesting to look at it like this @alexryzhkov! Thanks for pulling this together 🙂
\nI am the king of silver medals so it seems 😄 Would love to get gold either on co-visitation matrix - simplified, imprvd logic 🔥 or maybe even 💡 [Howto] Full dataset as parquet/csv files.
\nSo close arrrgh 😄
\nOh well, let's see how that will go 😄🍿
\nReally cool overview of the situation that you provide here! 🙌
", "rawMarkdown": "This is awesome stuff, wow! Really interesting to look at it like this @alexryzhkov! Thanks for pulling this together 🙂\n\nI am the king of silver medals so it seems 😄 Would love to get gold either on [co-visitation matrix - simplified, imprvd logic 🔥](https://www.kaggle.com/code/radek1/co-visitation-matrix-simplified-imprvd-logic) or maybe even [💡 [Howto] Full dataset as parquet/csv files](https://www.kaggle.com/code/radek1/howto-full-dataset-as-parquet-csv-files).\n\nSo close arrrgh 😄\n\nOh well, let's see how that will go 😄🍿\n\nReally cool overview of the situation that you provide here! 🙌", "votes": null }, { "id": "2023907", "postDate": "11/10/2022 06:04:08", "content": "Thanks @radek1 - now you are a litlle bit closer for golds :)
", "rawMarkdown": "Thanks @radek1 - now you are a litlle bit closer for golds :)", "votes": null } ], "comments": [ { "id": 2023724, "author_name": "radek1", "author_url": "", "post_date": "11/10/2022 01:32:49", "content": "This is awesome stuff, wow! Really interesting to look at it like this @alexryzhkov! Thanks for pulling this together 🙂
\nI am the king of silver medals so it seems 😄 Would love to get gold either on co-visitation matrix - simplified, imprvd logic 🔥 or maybe even 💡 [Howto] Full dataset as parquet/csv files.
\nSo close arrrgh 😄
\nOh well, let's see how that will go 😄🍿
\nReally cool overview of the situation that you provide here! 🙌
", "votes": null, "replies": [ { "id": 2023907, "author_name": "alexryzhkov", "author_url": "", "post_date": "11/10/2022 06:04:08", "content": "Thanks @radek1 - now you are a litlle bit closer for golds :)
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2023638": "Hi fellow Kagglers,\n\nI have created [the dataset](https://www.kaggle.com/datasets/alexryzhkov/otto-recommender-kernel-statistics) and [the example kernel](https://www.kaggle.com/code/alexryzhkov/otto-recommender-kernel-stats) to make some analytics on the kernels, which were created for this competition.\n\nIn the kernel you can see several interesting points like:\n- Top-20 kernel creators\n- Performance Tiers of kernel creators\n- Kernels medals stats\n- Medals for kernels creators (top-20)\n- Kernels languages\n- How many kernels are the forks?\n- How many lines are changed in forks?\n- Does the kernel use GPU or not?\n- Upvotes vs views and Upvotes vs PublicScore\n- Most commented kernels\n- **Kernel forks graph** and tops on it\n\nTo make it more interesting to check here are some results.\n\n### Medals for kernels\n\nHere we can congrats @cdeotte / @edwardcrookenden / @vslaykovsky / @simamumu who are dividing the current 1st place with 1 gold medal, @radek1 for the 2nd place with 4 silver medals and @ttahara for the 3rd place respectively with 2 silvers:\n\n\n\n### Forks graph\n\nThe full kernels graph for this competition looks like:\n\n\n\nI have used the kernel IDs in the graph to make it more readable so to find your kernel ID you can check it in the dataset by your name or kernel title.\n\nIf we use this graph and calculate the cumulative number of upvotes using the forks, we can receive the table like:\n\n\n\nThe current winner is @simamumu with his [[OLD] test_data last 20 `aid` get LB0.947](https://www.kaggle.com/code/simamumu/old-test-data-last-20-aid-get-lb0-947) kernel. @vslaykovsky is on the 2nd place and the 3rd is @cdeotte with amazing found - test data leak, which can help you to boost the LB score.\n\n### Kernel votes vs public score\n\nThe graph just says \"currently it's only the beginning of the competition - the score doesn't role so much now\" 🙃\n\n\n\nLooking forward for your comments, suggestions and insights!\n\nAlex", "2023724": "This is awesome stuff, wow! Really interesting to look at it like this @alexryzhkov! Thanks for pulling this together 🙂\n\nI am the king of silver medals so it seems 😄 Would love to get gold either on [co-visitation matrix - simplified, imprvd logic 🔥](https://www.kaggle.com/code/radek1/co-visitation-matrix-simplified-imprvd-logic) or maybe even [💡 [Howto] Full dataset as parquet/csv files](https://www.kaggle.com/code/radek1/howto-full-dataset-as-parquet-csv-files).\n\nSo close arrrgh 😄\n\nOh well, let's see how that will go 😄🍿\n\nReally cool overview of the situation that you provide here! 🙌", "2023907": "Thanks @radek1 - now you are a litlle bit closer for golds :)" }, "source": "meta" }