I saw a couple of great visualization like this one: https://www.kaggle.com/code/ivanrep/october-2022-tps-eda/notebook making a mistake interpreting the axis of the court (fixed from version 7).
\nIf we plot the trajectory of the ball in game_num=1, event_id=1002 filtering by the event_time that in the next 10s which one of the teams has scored, we observe that the last point is on the line y=-120.
\n ![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2F86d2547937b6d063f3d65dca51b9183a%2Fdownload1.png?generation=1664666848714186&alt=media\")
\nIf we repeat this with other events, we observe that the same thing happens (y=120 / y=-120):
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2F8e9ea43e6450d18ab9e6862f59928fd0%2Fdownload2.png?generation=1664666860777196&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2F34fb4f0e5b02a1cc9d46af00aeec9255%2Fdownload3.png?generation=1664666872021044&alt=media\")
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2Fcfc6636bf6ed6a5cbb475784a0a896fc%2Fdownload.png?generation=1664666988329599&alt=media\")
What does this mean?
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2Fe5294be952deabc6e74b3fce93dc811b%2Finbox_4696479_dd81bdc323e7fc146262b2272a35202f_WORNG.png?generation=1666104058035456&alt=media\")
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4696479%2Fc3be82731e4ba106e0faf8921784cf0a%2Finbox_981311_52acb29e8835b9248bc9b05f3170e9b1_2022-10-17-15-51-52-image.png?generation=1666104081770252&alt=media\")
I saw a couple of great visualization like this one: https://www.kaggle.com/code/ivanrep/october-2022-tps-eda/notebook making a mistake interpreting the axis of the court (fixed from version 7).
\nIf we plot the trajectory of the ball in game_num=1, event_id=1002 filtering by the event_time that in the next 10s which one of the teams has scored, we observe that the last point is on the line y=-120.
\n
\nIf we repeat this with other events, we observe that the same thing happens (y=120 / y=-120):
\n
\n
What does this mean?
\n
Thanks for noticing this. I'll make sure to redo this.
\nEdit: Fixed!
", "rawMarkdown": "Thanks for noticing this. I'll make sure to redo this.\n\nEdit: Fixed!", "votes": null }, { "id": "1966877", "postDate": "10/02/2022 07:48:11", "content": "Thanks to you for these amazing visualisations
", "rawMarkdown": "Thanks to you for these amazing visualisations", "votes": null }, { "id": "1967401", "postDate": "10/02/2022 13:35:46", "content": "@maherelouahabi there is nothing bad about transposing the field. it is not right or wrong, it is just a matter of taste. The important thing is to keep in mind a proper order of the coordinates when you inverse them. For example, I have explicitly pointed it out here.
", "rawMarkdown": "maherelouahabi there is nothing bad about transposing the field. it is not right or wrong, it is just a matter of taste. The important thing is to keep in mind a proper order of the coordinates when you inverse them. For example, [I have explicitly pointed it out here.](https://www.kaggle.com/code/sergiosaharovskiy/tps-oct-2022-viz-players-positions-animated)", "votes": null }, { "id": "1967472", "postDate": "10/02/2022 14:18:47", "content": "Nice Work!!
", "rawMarkdown": "Nice Work!!", "votes": null }, { "id": "1967514", "postDate": "10/02/2022 14:52:20", "content": "The problem is that it is not a transposition of the field. It is a transposition of the goals leaving the rest untransposed. This causes all the relative distances to the goal to be incorrect.
", "rawMarkdown": "The problem is that it is not a transposition of the field. It is a transposition of the goals leaving the rest untransposed. This causes all the relative distances to the goal to be incorrect.", "votes": null }, { "id": "1967589", "postDate": "10/02/2022 15:36:45", "content": "Lets say you have min X -82.5 and max X +82.5 and -121 Y, +121 Y respectively. You change them places and apply the same logic for the ball. It should not impact the visualization. Are you pointing out the specific work/s (notebook/s)?
", "rawMarkdown": "Lets say you have min X -82.5 and max X +82.5 and -121 Y, +121 Y respectively. You change them places and apply the same logic for the ball. It should not impact the visualization. Are you pointing out the specific work/s (notebook/s)?", "votes": null }, { "id": "1967617", "postDate": "10/02/2022 15:47:24", "content": "I totally agree with you. The key in what you said is 'apply the same logic for the ball'. NBs that I have observed, this logic wasn't extrapolated to the other coordinates such as the ball or the players. I thought it was important to point this out
", "rawMarkdown": "I totally agree with you. The key in what you said is 'apply the same logic for the ball'. NBs that I have observed, this logic wasn't extrapolated to the other coordinates such as the ball or the players. I thought it was important to point this out", "votes": null }, { "id": "1993607", "postDate": "10/18/2022 14:08:19", "content": "Hello Maher, first and foremost, thanks you for share this detail about the court. 😃
\nI wonder if according to what is shown in this code:
\ntrain.nlargest(1, 'ball_pos_x') #80.6875\ntrain.nsmallest(1, 'ball_pos_x') #-80.6250\ntrain.nlargest(1, 'ball_pos_y') #104.3125\ntrain.nsmallest(1, 'ball_pos_y')#-104.3125\nThis image fits better to what the X and Y axes indicate 😃
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F981311%2F52acb29e8835b9248bc9b05f3170e9b1%2F2022-10-17-15-51-52-image.png?generation=1666101937085772&alt=media\")
Kind Regards,
\nD.
", "rawMarkdown": "Hello Maher, first and foremost, thanks you for share this detail about the court. 😃\n\nI wonder if according to what is shown in this code:\n```\ntrain.nlargest(1, 'ball_pos_x') #80.6875\ntrain.nsmallest(1, 'ball_pos_x') #-80.6250\ntrain.nlargest(1, 'ball_pos_y') #104.3125\ntrain.nsmallest(1, 'ball_pos_y')#-104.3125\n```\n\nThis image fits better to what the X and Y axes indicate 😃\n\n\n\nKind Regards,\n\nD.", "votes": null }, { "id": "1993623", "postDate": "10/18/2022 14:32:38", "content": "Dear @davidhguerrero
\nYou are absolutely right. I was so blinded by showing the location of the goals that I misplaced the axes. Thank you for your correction. I will fix it.
\nUn saludo
\nMaher
Thanks for noticing this. I'll make sure to redo this.
\nEdit: Fixed!
", "votes": null, "replies": [ { "id": 1966877, "author_name": "maherelouahabi", "author_url": "", "post_date": "10/02/2022 07:48:11", "content": "Thanks to you for these amazing visualisations
", "votes": null, "replies": [] } ] }, { "id": 1967401, "author_name": "sergiosaharovskiy", "author_url": "", "post_date": "10/02/2022 13:35:46", "content": "@maherelouahabi there is nothing bad about transposing the field. it is not right or wrong, it is just a matter of taste. The important thing is to keep in mind a proper order of the coordinates when you inverse them. For example, I have explicitly pointed it out here.
", "votes": null, "replies": [ { "id": 1967514, "author_name": "maherelouahabi", "author_url": "", "post_date": "10/02/2022 14:52:20", "content": "The problem is that it is not a transposition of the field. It is a transposition of the goals leaving the rest untransposed. This causes all the relative distances to the goal to be incorrect.
", "votes": null, "replies": [] }, { "id": 1967589, "author_name": "sergiosaharovskiy", "author_url": "", "post_date": "10/02/2022 15:36:45", "content": "Lets say you have min X -82.5 and max X +82.5 and -121 Y, +121 Y respectively. You change them places and apply the same logic for the ball. It should not impact the visualization. Are you pointing out the specific work/s (notebook/s)?
", "votes": null, "replies": [] }, { "id": 1967617, "author_name": "maherelouahabi", "author_url": "", "post_date": "10/02/2022 15:47:24", "content": "I totally agree with you. The key in what you said is 'apply the same logic for the ball'. NBs that I have observed, this logic wasn't extrapolated to the other coordinates such as the ball or the players. I thought it was important to point this out
", "votes": null, "replies": [] } ] }, { "id": 1967472, "author_name": "aavaigmalhotra", "author_url": "", "post_date": "10/02/2022 14:18:47", "content": "Nice Work!!
", "votes": null, "replies": [] }, { "id": 1993607, "author_name": "davidhguerrero", "author_url": "", "post_date": "10/18/2022 14:08:19", "content": "Hello Maher, first and foremost, thanks you for share this detail about the court. 😃
\nI wonder if according to what is shown in this code:
\ntrain.nlargest(1, 'ball_pos_x') #80.6875\ntrain.nsmallest(1, 'ball_pos_x') #-80.6250\ntrain.nlargest(1, 'ball_pos_y') #104.3125\ntrain.nsmallest(1, 'ball_pos_y')#-104.3125\nThis image fits better to what the X and Y axes indicate 😃
\n
Kind Regards,
\nD.
", "votes": null, "replies": [ { "id": 1993623, "author_name": "maherelouahabi", "author_url": "", "post_date": "10/18/2022 14:32:38", "content": "Dear @davidhguerrero
\nYou are absolutely right. I was so blinded by showing the location of the goals that I misplaced the axes. Thank you for your correction. I will fix it.
\nUn saludo
\nMaher