{ "id": 395810, "title": "Has anyone managed to load the \"fullscreen\" data from the training and test datasets?", "url": "/competitions/predict-student-performance-from-game-play/discussion/395810", "author_name": "", "post_date": "2023-03-18T23:12:39.813888900Z", "votes": null, "comment_count": 8, "views": 0, "content": "

Has anyone been able to successfully load in the \"fullscreen\" column data from the test and train datasets?

\n

Everytime I try to load it in, it always just appears to be all \"na\"…

", "messages": [ { "id": "2187669", "postDate": "03/18/2023 23:12:39", "content": "

Has anyone been able to successfully load in the \"fullscreen\" column data from the test and train datasets?

\n

Everytime I try to load it in, it always just appears to be all \"na\"…

", "rawMarkdown": "Has anyone been able to successfully load in the \"fullscreen\" column data from the test and train datasets? \n\nEverytime I try to load it in, it always just appears to be all \"na\"...", "votes": null }, { "id": "2187675", "postDate": "03/18/2023 23:30:02", "content": "

The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in screen_coord_x and screen_coord_y.

", "rawMarkdown": "The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in `screen_coord_x` and `screen_coord_y`.", "votes": null }, { "id": "2187694", "postDate": "03/19/2023 00:25:20", "content": "
\n

The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in screen_coord_x and screen_coord_y.

\n
\n

Hey Chris, I'm sure this is also answered somewhere before. But are we supposed to populate the music and hq columns also in a similiar way since those are also NaN?

", "rawMarkdown": "> The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in `screen_coord_x` and `screen_coord_y`.\n\nHey Chris, I'm sure this is also answered somewhere before. But are we supposed to populate the music and hq columns also in a similiar way since those are also NaN?", "votes": null }, { "id": "2187709", "postDate": "03/19/2023 01:20:01", "content": "

The host said to ignore all 3 of these columns. There is no way to populate music or hq but we can populate fullscreen by analyzing the max screen x and y coords.

", "rawMarkdown": "The host said to ignore all 3 of these columns. There is no way to populate `music` or `hq` but we can populate `fullscreen` by analyzing the max screen x and y coords.", "votes": null }, { "id": "2188790", "postDate": "03/20/2023 01:46:57", "content": "

Silly question, what do you mean by the host?

", "rawMarkdown": "Silly question, what do you mean by the host?", "votes": null }, { "id": "2188802", "postDate": "03/20/2023 02:13:24", "content": "

The host is The Learning Agency Lab. This is their competition. They provided the data and asked Kaggle to run this competition. Somewhere in the discussions, they responded to the questions about NaN in these 3 columns. I searched quickly but cannot find their comment now.

", "rawMarkdown": "The host is `The Learning Agency Lab`. This is their competition. They provided the data and asked Kaggle to run this competition. Somewhere in the discussions, they responded to the questions about NaN in these 3 columns. I searched quickly but cannot find their comment now.", "votes": null }, { "id": "2188823", "postDate": "03/20/2023 02:50:40", "content": "

@cdeotte, when taking the maximum for screen_coord_xand screen_coord_y, you make two assumptions on the player, first is that he clicked on the far edges of his screen, second is that he didn't change his screen resolution along the way.
\nTo tackle this problem, you need to find a spot that the player has to click on, on a locked window, that is relatively small so that the error is the lowest possible, and observe for each value where that screen coordinate falls, something like this:
\n\"\"

\n

This is screen_coordinates of the click on one of the clues.
\nDots in the concentrated area are student who didn't change the standard resolution, and you can find the others by scaling it to this standard resolution.
\nI found that numerous sessions changed their screen resolution by trying to emulate game sessions. I invite you to consider this as it helped my model.

", "rawMarkdown": "cdeotte, when taking the maximum for `screen_coord_x `and `screen_coord_y`, you make two assumptions on the player, first is that he clicked on the far edges of his screen, second is that he didn't change his screen resolution along the way.\nTo tackle this problem, you need to find a spot that the player has to click on, on a locked window, that is relatively small so that the error is the lowest possible, and observe for each value where that screen coordinate falls, something like this: \n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F12870466%2Fa8cf155d043c0a5aee5f223daa87e1d5%2Ffor%20chris%20discussion.png?generation=1679280423803988&alt=media)\n\nThis is screen_coordinates of the click on one of the clues.\nDots in the concentrated area are student who didn't change the standard resolution, and you can find the others by scaling it to this standard resolution.\nI found that numerous sessions changed their screen resolution by trying to emulate game sessions. I invite you to consider this as it helped my model.", "votes": null }, { "id": "2195405", "postDate": "03/24/2023 16:03:47", "content": "

This is great platform for learner…i am enjoying here thanks for providing this community..

", "rawMarkdown": "This is great platform for learner...i am enjoying here thanks for providing this community..", "votes": null }, { "id": "2217456", "postDate": "04/10/2023 22:36:09", "content": "

I just used the max and min screen usage for the question session, I figured it more depended on what they actually did, not what the resolution was….

", "rawMarkdown": "I just used the max and min screen usage for the question session, I figured it more depended on what they actually did, not what the resolution was....", "votes": null } ], "comments": [ { "id": 2187675, "author_name": "cdeotte", "author_url": "", "post_date": "03/18/2023 23:30:02", "content": "

The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in screen_coord_x and screen_coord_y.

", "votes": null, "replies": [ { "id": 2187694, "author_name": "paulmerica", "author_url": "", "post_date": "03/19/2023 00:25:20", "content": "
\n

The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in screen_coord_x and screen_coord_y.

\n
\n

Hey Chris, I'm sure this is also answered somewhere before. But are we supposed to populate the music and hq columns also in a similiar way since those are also NaN?

", "votes": null, "replies": [ { "id": 2187709, "author_name": "cdeotte", "author_url": "", "post_date": "03/19/2023 01:20:01", "content": "

The host said to ignore all 3 of these columns. There is no way to populate music or hq but we can populate fullscreen by analyzing the max screen x and y coords.

", "votes": null, "replies": [ { "id": 2188790, "author_name": "gordyii", "author_url": "", "post_date": "03/20/2023 01:46:57", "content": "

Silly question, what do you mean by the host?

", "votes": null, "replies": [ { "id": 2188802, "author_name": "cdeotte", "author_url": "", "post_date": "03/20/2023 02:13:24", "content": "

The host is The Learning Agency Lab. This is their competition. They provided the data and asked Kaggle to run this competition. Somewhere in the discussions, they responded to the questions about NaN in these 3 columns. I searched quickly but cannot find their comment now.

", "votes": null, "replies": [] } ] }, { "id": 2188823, "author_name": "janmpia", "author_url": "", "post_date": "03/20/2023 02:50:40", "content": "

@cdeotte, when taking the maximum for screen_coord_xand screen_coord_y, you make two assumptions on the player, first is that he clicked on the far edges of his screen, second is that he didn't change his screen resolution along the way.
\nTo tackle this problem, you need to find a spot that the player has to click on, on a locked window, that is relatively small so that the error is the lowest possible, and observe for each value where that screen coordinate falls, something like this:
\n\"\"

\n

This is screen_coordinates of the click on one of the clues.
\nDots in the concentrated area are student who didn't change the standard resolution, and you can find the others by scaling it to this standard resolution.
\nI found that numerous sessions changed their screen resolution by trying to emulate game sessions. I invite you to consider this as it helped my model.

", "votes": null, "replies": [ { "id": 2217456, "author_name": "gordyii", "author_url": "", "post_date": "04/10/2023 22:36:09", "content": "

I just used the max and min screen usage for the question session, I figured it more depended on what they actually did, not what the resolution was….

", "votes": null, "replies": [] } ] } ] } ] } ] }, { "id": 2195405, "author_name": "prabinsharmaa", "author_url": "", "post_date": "03/24/2023 16:03:47", "content": "

This is great platform for learner…i am enjoying here thanks for providing this community..

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2187669": "Has anyone been able to successfully load in the \"fullscreen\" column data from the test and train datasets? \n\nEverytime I try to load it in, it always just appears to be all \"na\"...", "2187675": "The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in `screen_coord_x` and `screen_coord_y`.", "2187694": "> The host said that this column is all NaN so nobody can load it. You can create your own column by checking the max values in `screen_coord_x` and `screen_coord_y`.\n\nHey Chris, I'm sure this is also answered somewhere before. But are we supposed to populate the music and hq columns also in a similiar way since those are also NaN?", "2187709": "The host said to ignore all 3 of these columns. There is no way to populate `music` or `hq` but we can populate `fullscreen` by analyzing the max screen x and y coords.", "2188790": "Silly question, what do you mean by the host?", "2188802": "The host is `The Learning Agency Lab`. This is their competition. They provided the data and asked Kaggle to run this competition. Somewhere in the discussions, they responded to the questions about NaN in these 3 columns. I searched quickly but cannot find their comment now.", "2188823": "cdeotte, when taking the maximum for `screen_coord_x `and `screen_coord_y`, you make two assumptions on the player, first is that he clicked on the far edges of his screen, second is that he didn't change his screen resolution along the way.\nTo tackle this problem, you need to find a spot that the player has to click on, on a locked window, that is relatively small so that the error is the lowest possible, and observe for each value where that screen coordinate falls, something like this: \n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F12870466%2Fa8cf155d043c0a5aee5f223daa87e1d5%2Ffor%20chris%20discussion.png?generation=1679280423803988&alt=media)\n\nThis is screen_coordinates of the click on one of the clues.\nDots in the concentrated area are student who didn't change the standard resolution, and you can find the others by scaling it to this standard resolution.\nI found that numerous sessions changed their screen resolution by trying to emulate game sessions. I invite you to consider this as it helped my model.", "2195405": "This is great platform for learner...i am enjoying here thanks for providing this community..", "2217456": "I just used the max and min screen usage for the question session, I figured it more depended on what they actually did, not what the resolution was...." }, "source": "meta" }