{ "id": 387864, "title": "How The Game Begins - EDA", "url": "/competitions/predict-student-performance-from-game-play/discussion/387864", "author_name": "Chris Deotte", "post_date": "2023-02-15T00:32:31.908000", "votes": 228, "comment_count": 58, "views": 0, "content": "

How The Game Begins - EDA

\n

When the game begins, our web browser pops out a new window of size 660x880. The following 11 screen shots are the first 11 things we see. We do not have a choice. The game pauses on each image until we click anywhere.

\n

The first screen we see is the image in the top left (which is also image in yellow box in room tunic.historicalsociety.closet below). No matter where we click, grampa says \"Whatcha doing over there, Jo?\". Then no matter where we click, we say \"Just talking to Teddy.\". The dialog continues as pictured below. (Right click the images below and choose \"Open Image in New Tab\" to enlarge images).

\n

\"\"

\n

First Game Room - tunic.historicalsociety.closet

\n

Each row of the dataset has 2 pairs of coordinates, (room_coor_x, room_coor_y) and (screen_coor_x, screen_coor_y). The first pair is relative to the black rectangle below. And the second pair is relative to your pop out window (not your computer screen). Two examples of pop up window views are the yellow and blue rectanges below.

\n

In the room tunic.historicalsociety.closet (i.e. the first room), then -1064<=room_coor_x<=926 and -393<=room_coor_y<=327. Each room has its own bounds. The screen coordinates are always bound by your pop out window size. 95% of users do not resize the pop out window and 0<=room_coor_x<=880 and 0<=room_coor_y<=660

\n

\"\"

\n

First Click Option

\n

When the game begins, we see the image in the yellow rectangle above. No matter where we click, grampa will say \"Whatcha doing over there, Jo?\", but it is still interesting to see where users click first. Some users click grampa. Some users click the retirement_letter on grampa's desk. See the clusters of points below. Each blue dot is a different users first click:

\n

Room Coordinates

\n

We plot plt.scatter(df.room_coor_x, df.room_coor_y) in the black box:
\n\"\"

\n

Screen Coordinates

\n

We plot plt.scatter(df.screen_coor_x, 660-df.screen_coor_y) in the yellow box. Note we flip screen_coor_y:
\n\"\"

\n

First Action Option

\n

After we finish the mandatory introductory dialog, we are free to choose our first action. The scatter plot below shows where users make their first action. Most users click Teddy (the lovable pet badger). Some (curious users) click the photos on the wall. Some click to the left of the room to navigate (eager to complete their first quest):

\n

\"\"

\n

First Quest

\n

The last message in the dialog (before Teddy's emoji) is Jo saying, \"Now where did I put my notebook?\". This is our first quest. We must find the notebook. If we traverse to the far left side of this first room, we see the notebook on the floor. Every user must find and click this notebook to proceed. Below are every users first click finding the notebook. When we hover over the notebook, we see an orange exclamation point. Some users click the exclamation point, and some users click directly on the notebook:

\n

\"\"

\n

First Use of Notebook

\n

After clicking the notebook, the notebook opens and takes over the full pop up window. The message \"Found it!\" appears in the bottom of the screen over the notebook. To continue the game, we can click anywhere. Most users click on the words \"Found it\", but some users click elsewhere. Regardless of where we click, the message disappears and a white X in black circle appears in top right corner. Most users click the white X in black circle which closes the notebook and play continues. Some users click elsewhere and nothing happens. They must keep clicking until they click the white X in black circle. Some users even click the words \"Homework Monster\" in the bottom center of the notebook!

\n

\"\"
\nWe plot plt.scatter(df.screen_coor_x, 660-df.screen_coor_y). Note we flip screen_coor_y:
\n\"\"

\n

User Behavior - Feature Engineering

\n

There is so much to learn about the 11k users by just observing their behavior in this opening sequence. How long do they take to read each text dialog before continuing? Where do they click each time to continue the game. Where is their first click (grampa, letter, elsewhere)? When users click grampa, do they click grampa's body or face? When Teddy appears on the screen, how many users start clicking the pet versus clicking on Jo to continue the dialog? Which users click text boxes to continue. Which users resize the initial pop up window, which don't? etc etc.

\n

This is just the opening sequence. There are 19 rooms and 23 levels to explore. Each moment, users take one of 11 actions. There are many dialogs and many quests! Different users click different things. Users reveal much predictive information (i.e. features) by their game playing behavior!

\n

UPDATE - Starter Notebook

\n

I published a starter notebook here displaying click EDA for all 19 game rooms. The starter notebook displays navigate_clicks and person_clicks and labels items of interest in each room

", "messages": [ { "id": 2145327, "postDate": "2023-02-15T00:32:31.910Z", "content": "

How The Game Begins - EDA

\n

When the game begins, our web browser pops out a new window of size 660x880. The following 11 screen shots are the first 11 things we see. We do not have a choice. The game pauses on each image until we click anywhere.

\n

The first screen we see is the image in the top left (which is also image in yellow box in room tunic.historicalsociety.closet below). No matter where we click, grampa says \"Whatcha doing over there, Jo?\". Then no matter where we click, we say \"Just talking to Teddy.\". The dialog continues as pictured below. (Right click the images below and choose \"Open Image in New Tab\" to enlarge images).

\n

\"\"

\n

First Game Room - tunic.historicalsociety.closet

\n

Each row of the dataset has 2 pairs of coordinates, (room_coor_x, room_coor_y) and (screen_coor_x, screen_coor_y). The first pair is relative to the black rectangle below. And the second pair is relative to your pop out window (not your computer screen). Two examples of pop up window views are the yellow and blue rectanges below.

\n

In the room tunic.historicalsociety.closet (i.e. the first room), then -1064<=room_coor_x<=926 and -393<=room_coor_y<=327. Each room has its own bounds. The screen coordinates are always bound by your pop out window size. 95% of users do not resize the pop out window and 0<=room_coor_x<=880 and 0<=room_coor_y<=660

\n

\"\"

\n

First Click Option

\n

When the game begins, we see the image in the yellow rectangle above. No matter where we click, grampa will say \"Whatcha doing over there, Jo?\", but it is still interesting to see where users click first. Some users click grampa. Some users click the retirement_letter on grampa's desk. See the clusters of points below. Each blue dot is a different users first click:

\n

Room Coordinates

\n

We plot plt.scatter(df.room_coor_x, df.room_coor_y) in the black box:
\n\"\"

\n

Screen Coordinates

\n

We plot plt.scatter(df.screen_coor_x, 660-df.screen_coor_y) in the yellow box. Note we flip screen_coor_y:
\n\"\"

\n

First Action Option

\n

After we finish the mandatory introductory dialog, we are free to choose our first action. The scatter plot below shows where users make their first action. Most users click Teddy (the lovable pet badger). Some (curious users) click the photos on the wall. Some click to the left of the room to navigate (eager to complete their first quest):

\n

\"\"

\n

First Quest

\n

The last message in the dialog (before Teddy's emoji) is Jo saying, \"Now where did I put my notebook?\". This is our first quest. We must find the notebook. If we traverse to the far left side of this first room, we see the notebook on the floor. Every user must find and click this notebook to proceed. Below are every users first click finding the notebook. When we hover over the notebook, we see an orange exclamation point. Some users click the exclamation point, and some users click directly on the notebook:

\n

\"\"

\n

First Use of Notebook

\n

After clicking the notebook, the notebook opens and takes over the full pop up window. The message \"Found it!\" appears in the bottom of the screen over the notebook. To continue the game, we can click anywhere. Most users click on the words \"Found it\", but some users click elsewhere. Regardless of where we click, the message disappears and a white X in black circle appears in top right corner. Most users click the white X in black circle which closes the notebook and play continues. Some users click elsewhere and nothing happens. They must keep clicking until they click the white X in black circle. Some users even click the words \"Homework Monster\" in the bottom center of the notebook!

\n

\"\"
\nWe plot plt.scatter(df.screen_coor_x, 660-df.screen_coor_y). Note we flip screen_coor_y:
\n\"\"

\n

User Behavior - Feature Engineering

\n

There is so much to learn about the 11k users by just observing their behavior in this opening sequence. How long do they take to read each text dialog before continuing? Where do they click each time to continue the game. Where is their first click (grampa, letter, elsewhere)? When users click grampa, do they click grampa's body or face? When Teddy appears on the screen, how many users start clicking the pet versus clicking on Jo to continue the dialog? Which users click text boxes to continue. Which users resize the initial pop up window, which don't? etc etc.

\n

This is just the opening sequence. There are 19 rooms and 23 levels to explore. Each moment, users take one of 11 actions. There are many dialogs and many quests! Different users click different things. Users reveal much predictive information (i.e. features) by their game playing behavior!

\n

UPDATE - Starter Notebook

\n

I published a starter notebook here displaying click EDA for all 19 game rooms. The starter notebook displays navigate_clicks and person_clicks and labels items of interest in each room

", "rawMarkdown": "# How The Game Begins - EDA\nWhen the game begins, our web browser pops out a new window of size 660x880. The following 11 screen shots are the first 11 things we see. We do not have a choice. The game pauses on each image until we click anywhere. \n\nThe first screen we see is the image in the top left (which is also image in yellow box in room tunic.historicalsociety.closet below). No matter where we click, grampa says \"Whatcha doing over there, Jo?\". Then no matter where we click, we say \"Just talking to Teddy.\". The dialog continues as pictured below. (Right click the images below and choose \"Open Image in New Tab\" to enlarge images).\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c12.png)\n\n# First Game Room - tunic.historicalsociety.closet\nEach row of the dataset has 2 pairs of coordinates, `(room_coor_x, room_coor_y)` and `(screen_coor_x, screen_coor_y)`. The first pair is relative to the black rectangle below. And the second pair is relative to your pop out window (not your computer screen). Two examples of pop up window views are the yellow and blue rectanges below. \n\nIn the room tunic.historicalsociety.closet (i.e. the first room), then `-1064<=room_coor_x<=926` and `-393<=room_coor_y<=327`. Each room has its own bounds. The screen coordinates are always bound by your pop out window size. 95% of users do not resize the pop out window and `0<=room_coor_x<=880` and `0<=room_coor_y<=660`\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c2.png)\n\n# First Click Option\nWhen the game begins, we see the image in the yellow rectangle above. No matter where we click, grampa will say \"Whatcha doing over there, Jo?\", but it is still interesting to see where users click first. Some users click `grampa`. Some users click the `retirement_letter` on grampa's desk. See the clusters of points below. Each blue dot is a different users first click:\n### Room Coordinates\nWe plot `plt.scatter(df.room_coor_x, df.room_coor_y)` in the black box:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c3.png)\n### Screen Coordinates\nWe plot `plt.scatter(df.screen_coor_x, 660-df.screen_coor_y)` in the yellow box. Note we flip `screen_coor_y`:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c4.png)\n\n# First Action Option\nAfter we finish the mandatory introductory dialog, we are free to choose our first action. The scatter plot below shows where users make their first action. Most users click Teddy (the lovable pet badger). Some (curious users) click the photos on the wall. Some click to the left of the room to navigate (eager to complete their first quest):\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c5.png)\n\n# First Quest\nThe last message in the dialog (before Teddy's emoji) is Jo saying, \"Now where did I put my notebook?\". This is our first quest. We must find the notebook. If we traverse to the far left side of this first room, we see the notebook on the floor. Every user must find and click this notebook to proceed. Below are every users first click finding the notebook. When we hover over the notebook, we see an orange exclamation point. Some users click the exclamation point, and some users click directly on the notebook:\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c7.png)\n\n# First Use of Notebook\nAfter clicking the notebook, the notebook opens and takes over the full pop up window. The message \"Found it!\" appears in the bottom of the screen over the notebook. To continue the game, we can click anywhere. Most users click on the words \"Found it\", but some users click elsewhere. Regardless of where we click, the message disappears and a white X in black circle appears in top right corner. Most users click the white X in black circle which closes the notebook and play continues. Some users click elsewhere and nothing happens. They must keep clicking until they click the white X in black circle. Some users even click the words \"Homework Monster\" in the bottom center of the notebook!\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c8.png)\nWe plot `plt.scatter(df.screen_coor_x, 660-df.screen_coor_y)`. Note we flip `screen_coor_y`:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c9.png)\n\n# User Behavior - Feature Engineering\nThere is so much to learn about the 11k users by just observing their behavior in this opening sequence. How long do they take to read each text dialog before continuing? Where do they click each time to continue the game. Where is their first click (grampa, letter, elsewhere)? When users click grampa, do they click grampa's body or face? When Teddy appears on the screen, how many users start clicking the pet versus clicking on Jo to continue the dialog? Which users click text boxes to continue. Which users resize the initial pop up window, which don't? etc etc.\n\nThis is just the opening sequence. There are 19 rooms and 23 levels to explore. Each moment, users take one of 11 actions. There are many dialogs and many quests! Different users click different things. Users reveal much predictive information (i.e. features) by their game playing behavior!\n\n# UPDATE - Starter Notebook\nI published a starter notebook [here][1] displaying click EDA for all 19 game rooms. The starter notebook displays `navigate_clicks` and `person_clicks` and labels items of interest in each room\n\n[1]: https://www.kaggle.com/code/cdeotte/game-room-click-eda", "votes": 228 }, { "id": 2145601, "postDate": "2023-02-15T08:19:59.113Z", "content": "

Thank you ! This is informative and well illustrated.

\n

Also, it's been a long time that I wanted to express that : huge THANK YOU for all your efforts and sharings with the community ! And not only on this competition 🙏

", "rawMarkdown": "Thank you ! This is informative and well illustrated.\n\nAlso, it's been a long time that I wanted to express that : huge THANK YOU for all your efforts and sharings with the community ! And not only on this competition 🙏", "votes": 3 }, { "id": 2145608, "postDate": "2023-02-15T08:29:33.483Z", "content": "

UPDATE - Starter Notebook
\nI published a starter notebook here displaying click EDA for all 19 game rooms. The starter notebook displays navigate_clicks and person_clicks and labels items of interest in each room

", "rawMarkdown": "**UPDATE - Starter Notebook**\nI published a starter notebook [here][1] displaying click EDA for all 19 game rooms. The starter notebook displays `navigate_clicks` and `person_clicks` and labels items of interest in each room\n\n[1]: https://www.kaggle.com/cdeotte/game-room-click-eda", "votes": 4 }, { "id": 2145526, "postDate": "2023-02-15T06:51:41.663Z", "content": "

I make lots of features based on diff 'elapsed_time', which is really useful. But I have long way to go. Thanks for your impressive illustration.

", "rawMarkdown": "I make lots of features based on diff 'elapsed_time', which is really useful. But I have long way to go. Thanks for your impressive illustration.", "votes": 2 }, { "id": 2218126, "postDate": "2023-04-11T13:05:23.753Z", "content": "

Very helpful!

", "rawMarkdown": "Very helpful!", "votes": 1 }, { "id": 2203674, "postDate": "2023-03-31T03:18:43.487Z", "content": "

Interesting!

", "rawMarkdown": "Interesting!", "votes": 1 }, { "id": 2196424, "postDate": "2023-03-25T11:56:26.187Z", "content": "

It's Brilliant

", "rawMarkdown": "It's Brilliant", "votes": 1 }, { "id": 2185294, "postDate": "2023-03-16T23:40:06.550Z", "content": "

Very helpful!

", "rawMarkdown": "Very helpful!", "votes": 1 }, { "id": 2185254, "postDate": "2023-03-16T22:45:12.503Z", "content": "

What an amazing insight! Just for my curiosity, let me know how did you get it?

", "rawMarkdown": "What an amazing insight! Just for my curiosity, let me know how did you get it?", "votes": 1 }, { "id": 2159116, "postDate": "2023-02-25T13:02:39.393Z", "content": "

Interesting! I actually played the game. It was about the same click.

", "rawMarkdown": "Interesting! I actually played the game. It was about the same click.", "votes": 1 }, { "id": 2158497, "postDate": "2023-02-24T21:58:36.897Z", "content": "

Nice and helpful !

", "rawMarkdown": "Nice and helpful !", "votes": 1 }, { "id": 2156350, "postDate": "2023-02-23T09:33:44.127Z", "content": "

Interesting

", "rawMarkdown": "Interesting", "votes": 1 }, { "id": 2155324, "postDate": "2023-02-22T15:12:22.443Z", "content": "

helpful thx

", "rawMarkdown": "helpful thx ", "votes": 1 }, { "id": 2155104, "postDate": "2023-02-22T13:01:57.810Z", "content": "

very helpful！

", "rawMarkdown": "very helpful！", "votes": 1 }, { "id": 2154676, "postDate": "2023-02-22T07:25:27.450Z", "content": "

Thanks for sharing. I hope this will help to understand EDA.

", "rawMarkdown": "Thanks for sharing. I hope this will help to understand EDA.", "votes": 1 }, { "id": 2154511, "postDate": "2023-02-22T04:30:17.810Z", "content": "

Helpfull thx

", "rawMarkdown": "Helpfull thx", "votes": 1 }, { "id": 2154377, "postDate": "2023-02-22T01:14:21.673Z", "content": "

very helpful!

", "rawMarkdown": "very helpful!", "votes": 1 }, { "id": 2154087, "postDate": "2023-02-21T19:59:03.803Z", "content": "

Nice notebook

", "rawMarkdown": "Nice notebook", "votes": 1 }, { "id": 2153902, "postDate": "2023-02-21T17:37:03.297Z", "content": "

very helpful

", "rawMarkdown": "very helpful", "votes": 1 }, { "id": 2150307, "postDate": "2023-02-19T06:22:16.997Z", "content": "

Very helpful!

", "rawMarkdown": "Very helpful!", "votes": 1 }, { "id": 2149767, "postDate": "2023-02-18T16:48:39.303Z", "content": "

Interesting!

", "rawMarkdown": "Interesting!", "votes": 1 }, { "id": 2149098, "postDate": "2023-02-18T00:07:23.963Z", "content": "

Interesting!

", "rawMarkdown": "Interesting!", "votes": 1 }, { "id": 2148523, "postDate": "2023-02-17T14:10:07.547Z", "content": "

Interesting!

", "rawMarkdown": "Interesting!", "votes": 1 }, { "id": 2148492, "postDate": "2023-02-17T13:28:25.893Z", "content": "

Interesting!

", "rawMarkdown": "Interesting!", "votes": 1 }, { "id": 2147929, "postDate": "2023-02-17T02:50:58.993Z", "content": "

This analysis is interesting! It makes me enthusiastic. Thank you.

", "rawMarkdown": "This analysis is interesting! It makes me enthusiastic. Thank you.", "votes": 1 }, { "id": 2147898, "postDate": "2023-02-17T01:10:45.573Z", "content": "

This insightful analysis helps me a lot! Thank you.

", "rawMarkdown": "This insightful analysis helps me a lot! Thank you.", "votes": 1 }, { "id": 2147623, "postDate": "2023-02-16T18:38:54.513Z", "content": "

Helpfull thx

", "rawMarkdown": "Helpfull thx", "votes": 1 }, { "id": 2147573, "postDate": "2023-02-16T18:01:42.937Z", "content": "

@cdeotte , you are amazing….

", "rawMarkdown": "@cdeotte , you are amazing.... ", "votes": 1 }, { "id": 2146644, "postDate": "2023-02-16T03:08:01.480Z", "content": "

@cdeotte awesome notebook and visualization!

", "rawMarkdown": "@cdeotte awesome notebook and visualization!", "votes": 1 }, { "id": 2146096, "postDate": "2023-02-15T15:50:51.250Z", "content": "

Thank your Chris for the extra analysis.It seems that this competition is all but a game…

", "rawMarkdown": "Thank your Chris for the extra analysis.It seems that this competition is all but a game...", "votes": 1 }, { "id": 2145734, "postDate": "2023-02-15T10:09:38.260Z", "content": "

Thanks for sharing this useful information, Chris

", "rawMarkdown": "Thanks for sharing this useful information, Chris", "votes": 1 }, { "id": 2187966, "postDate": "2023-03-19T07:33:07.773Z", "content": "

Very Interesting!

", "rawMarkdown": "Very Interesting!", "votes": 2 }, { "id": 2186009, "postDate": "2023-03-17T13:45:55.050Z", "content": "

Nice notebook

", "rawMarkdown": "Nice notebook", "votes": 2 }, { "id": 2200230, "postDate": "2023-03-28T11:49:56.523Z", "content": "

Very Interesting!

", "rawMarkdown": "Very Interesting!" }, { "id": 2147749, "postDate": "2023-02-16T20:59:47.023Z", "content": "

I found that about half of all sessions have 4 screens different: \"Can I come, Gramps?\" and related 3 more lines of dialogue instead of \"Meetings are BORING!\" and related 3 more lines of of dialogue. There's even some with \"Ugh. Meetings are so boring.\"

\n

But when I try a session, I always get \"Meetings are BORING!\". If there's a place to click or a secret, I haven't found it yet. EDA didn't show any relevant difference in clicks. There was a big difference in average 'elapsed time'(!) but when I try fast clicking through I didn't get any different result in my personal sessions.

", "rawMarkdown": "I found that about half of all sessions have 4 screens different: \"Can I come, Gramps?\" and related 3 more lines of dialogue instead of \"Meetings are BORING!\" and related 3 more lines of of dialogue. There's even some with \"Ugh. Meetings are so boring.\"\n\nBut when I try a session, I always get \"Meetings are BORING!\". If there's a place to click or a secret, I haven't found it yet. EDA didn't show any relevant difference in clicks. There was a big difference in average 'elapsed time'(!) but when I try fast clicking through I didn't get any different result in my personal sessions.\n\n", "votes": 2, "replies": [ { "id": 2147760, "postDate": "2023-02-16T21:22:13.833Z", "content": "

There are four different scripts for the game that are randomly chosen. Got this from the Jo Wilder GitHub repo https://github.com/fielddaylab/jo_wilder :

\n

Script Types

\n

Starting in v7, multiple scripts were added to the game for AB tests on snark and humor. The game will randomly choose between 4 different data files: data_dry.js, data_nohumor.js, data_nosnark.js, and the original data.js. They are each built from their respective data folder in assets. The type of script used is only logged once, in startgame.

\n

Index Name Description
\n0 dry no humor or snark
\n1 nohumor no humor (includes snark)
\n2 nosnark no snark (includes humor). No snark can also be thought of as \"obedient\"
\n3 normal base script (includes snark and humor)

", "rawMarkdown": "There are four different scripts for the game that are randomly chosen. Got this from the Jo Wilder GitHub repo [https://github.com/fielddaylab/jo_wilder](url) :\n\nScript Types\n\nStarting in v7, multiple scripts were added to the game for AB tests on snark and humor. The game will randomly choose between 4 different data files: data_dry.js, data_nohumor.js, data_nosnark.js, and the original data.js. They are each built from their respective data folder in assets. The type of script used is only logged once, in startgame.\n\nIndex\tName\tDescription\n0\tdry\tno humor or snark\n1\tnohumor\tno humor (includes snark)\n2\tnosnark\tno snark (includes humor). No snark can also be thought of as \"obedient\"\n3\tnormal\tbase script (includes snark and humor)", "votes": 9 }, { "id": 2147776, "postDate": "2023-02-16T21:48:54.887Z", "content": "

Yes. In the training data not all users are playing the same game events. Therefore we need to be creative at times to find the same thing in all users history.

", "rawMarkdown": "Yes. In the training data not all users are playing the same game events. Therefore we need to be creative at times to find the same thing in all users history.", "votes": 1 }, { "id": 2147780, "postDate": "2023-02-16T21:51:26.190Z", "content": "

For example in my new EDA post here. We cannot use text to locate these events in all users. We must use other features to find these events in all users. There are things that are common like level and fqid etc etc to match events between all users. The biggest difference i noticed was different text in different sessions.

", "rawMarkdown": "For example in my new EDA post [here][1]. We cannot use text to locate these events in all users. We must use other features to find these events in all users. There are things that are common like `level` and `fqid` etc etc to match events between all users. The biggest difference i noticed was different text in different sessions.\n\n[1]: https://www.kaggle.com/competitions/predict-student-performance-from-game-play/discussion/388094", "votes": 3 } ] }, { "id": 2146354, "postDate": "2023-02-15T19:59:12.943Z", "content": "

This is really helpful stuff, Thanks @cdeotte

", "rawMarkdown": "This is really helpful stuff, Thanks @cdeotte ", "votes": 2 }, { "id": 2145622, "postDate": "2023-02-15T08:50:21.647Z", "content": "

Thank you Chris for this awesome analysis!

", "rawMarkdown": "Thank you Chris for this awesome analysis! ", "votes": 2 }, { "id": 2292271, "postDate": "2023-06-08T07:14:41.593Z", "content": "

Good insights ! Thx!

", "rawMarkdown": "Good insights ! Thx!" }, { "id": 2232779, "postDate": "2023-04-24T15:50:11.570Z", "content": "

Your sharing is always so helpful！!

", "rawMarkdown": "Your sharing is always so helpful！!" }, { "id": 2219297, "postDate": "2023-04-12T13:23:31.143Z", "content": "

Interesting! @cdeotte
\nHow did you know it was that screen?

", "rawMarkdown": "Interesting! @cdeotte\nHow did you know it was that screen? " }, { "id": 2269416, "postDate": "2023-05-22T13:16:11.370Z", "rawMarkdown": "", "isDeleted": true }, { "id": 2202984, "postDate": "2023-03-30T12:45:35.453Z", "content": "

thank you very much

", "rawMarkdown": "thank you very much", "votes": 1 }, { "id": 2202656, "postDate": "2023-03-30T08:02:03.773Z", "content": "

Thanks a lot!👍

", "rawMarkdown": "Thanks a lot!👍", "votes": 1 }, { "id": 2184740, "postDate": "2023-03-16T15:38:03.980Z", "content": "

Very impressed! Thanks for sharing!

", "rawMarkdown": "Very impressed! Thanks for sharing!", "votes": 1 }, { "id": 2184428, "postDate": "2023-03-16T11:29:49.807Z", "content": "

Thanks, it is helpful!

", "rawMarkdown": "Thanks, it is helpful!", "votes": 1 }, { "id": 2183633, "postDate": "2023-03-15T19:36:50.257Z", "content": "

Very informative. Thanks!

", "rawMarkdown": "Very informative. Thanks!", "votes": 1 }, { "id": 2183055, "postDate": "2023-03-15T12:58:26.987Z", "content": "

Thanks a lot XD

", "rawMarkdown": "Thanks a lot XD", "votes": 1 }, { "id": 2157203, "postDate": "2023-02-23T19:26:16.863Z", "content": "

Thanks for your explanation

", "rawMarkdown": "Thanks for your explanation", "votes": 1 }, { "id": 2156642, "postDate": "2023-02-23T13:01:31.690Z", "content": "

Thanks for sharing. very helpful!

", "rawMarkdown": "Thanks for sharing. very helpful!", "votes": 1 }, { "id": 2155857, "postDate": "2023-02-23T00:33:27.993Z", "content": "

Thanks for this

", "rawMarkdown": "Thanks for this", "votes": 1 }, { "id": 2151295, "postDate": "2023-02-20T02:04:16.837Z", "content": "

very helpful, thanks for sharing!

", "rawMarkdown": "very helpful, thanks for sharing!", "votes": 1 }, { "id": 2149210, "postDate": "2023-02-18T04:33:40.670Z", "content": "

It's really helpful! Thank you!

", "rawMarkdown": "It's really helpful! Thank you!", "votes": 1 }, { "id": 2147709, "postDate": "2023-02-16T20:11:18.327Z", "content": "

very helpful thanks!

", "rawMarkdown": "very helpful thanks!", "votes": 1 }, { "id": 2147357, "postDate": "2023-02-16T15:03:29.097Z", "content": "

Thanks for sharing.

", "rawMarkdown": "Thanks for sharing.", "votes": 1 }, { "id": 2147032, "postDate": "2023-02-16T11:16:21.163Z", "content": "

Thanks deotte

", "rawMarkdown": "Thanks deotte", "votes": 1 }, { "id": 2145499, "postDate": "2023-02-15T06:23:06.970Z", "content": "

Awesome! Thanks a lot Chris.

", "rawMarkdown": "Awesome! Thanks a lot Chris.", "votes": 1 } ], "comments": [ { "id": 2145601, "author_name": "Vincent Schuler", "author_url": "", "post_date": "2023-02-15T08:19:59.113000", "content": "

Thank you ! This is informative and well illustrated.

\n

Also, it's been a long time that I wanted to express that : huge THANK YOU for all your efforts and sharings with the community ! And not only on this competition 🙏

", "votes": 3, "replies": [] }, { "id": 2145608, "author_name": "Chris Deotte", "author_url": "", "post_date": "2023-02-15T08:29:33.483000", "content": "

UPDATE - Starter Notebook
\nI published a starter notebook here displaying click EDA for all 19 game rooms. The starter notebook displays navigate_clicks and person_clicks and labels items of interest in each room

", "votes": 4, "replies": [] }, { "id": 2145526, "author_name": "Joseph Zhou", "author_url": "", "post_date": "2023-02-15T06:51:41.663000", "content": "

I make lots of features based on diff 'elapsed_time', which is really useful. But I have long way to go. Thanks for your impressive illustration.

", "votes": 2, "replies": [] }, { "id": 2218126, "author_name": "Byungeun Hwang", "author_url": "", "post_date": "2023-04-11T13:05:23.753000", "content": "

Very helpful!

", "votes": 1, "replies": [] }, { "id": 2203674, "author_name": "hagdon", "author_url": "", "post_date": "2023-03-31T03:18:43.487000", "content": "

Interesting!

", "votes": 1, "replies": [] }, { "id": 2196424, "author_name": "Alex Skuratov", "author_url": "", "post_date": "2023-03-25T11:56:26.187000", "content": "

It's Brilliant

", "votes": 1, "replies": [] }, { "id": 2185294, "author_name": "takekawa tomoki", "author_url": "", "post_date": "2023-03-16T23:40:06.550000", "content": "

Very helpful!

", "votes": 1, "replies": [] }, { "id": 2185254, "author_name": "JumpeiT", "author_url": "", "post_date": "2023-03-16T22:45:12.503000", "content": "

What an amazing insight! Just for my curiosity, let me know how did you get it?

", "votes": 1, "replies": [] }, { "id": 2159116, "author_name": "knn180", "author_url": "", "post_date": "2023-02-25T13:02:39.393000", "content": "

Interesting! I actually played the game. It was about the same click.

", "votes": 1, "replies": [] }, { "id": 2158497, "author_name": "Sitraka_Forler", "author_url": "", "post_date": "2023-02-24T21:58:36.897000", "content": "

Nice and helpful !

", "votes": 1, "replies": [] }, { "id": 2156350, "author_name": "Shazia Saif", "author_url": "", "post_date": "2023-02-23T09:33:44.127000", "content": "

Interesting

", "votes": 1, "replies": [] }, { "id": 2155324, "author_name": "zhengtianl", "author_url": "", "post_date": "2023-02-22T15:12:22.443000", "content": "

helpful thx

", "votes": 1, "replies": [] }, { "id": 2155104, "author_name": "empKiros", "author_url": "", "post_date": "2023-02-22T13:01:57.810000", "content": "

very helpful！

", "votes": 1, "replies": [] }, { "id": 2154676, "author_name": "Muhammad Usama", "author_url": "", "post_date": "2023-02-22T07:25:27.450000", "content": "

Thanks for sharing. I hope this will help to understand EDA.

", "votes": 1, "replies": [] }, { "id": 2154511, "author_name": "reubenpofb", "author_url": "", "post_date": "2023-02-22T04:30:17.810000", "content": "

Helpfull thx

", "votes": 1, "replies": [] }, { "id": 2154377, "author_name": "Weizi Qian", "author_url": "", "post_date": "2023-02-22T01:14:21.673000", "content": "

very helpful!

", "votes": 1, "replies": [] }, { "id": 2154087, "author_name": "max phillep", "author_url": "", "post_date": "2023-02-21T19:59:03.803000", "content": "

Nice notebook

", "votes": 1, "replies": [] }, { "id": 2153902, "author_name": "Bunny", "author_url": "", "post_date": "2023-02-21T17:37:03.297000", "content": "

very helpful

", "votes": 1, "replies": [] }, { "id": 2150307, "author_name": "Lu Gan", "author_url": "", "post_date": "2023-02-19T06:22:16.997000", "content": "

Very helpful!

", "votes": 1, "replies": [] }, { "id": 2149767, "author_name": "Wided Zgouji", "author_url": "", "post_date": "2023-02-18T16:48:39.303000", "content": "

Interesting!

", "votes": 1, "replies": [] }, { "id": 2149098, "author_name": "1025107", "author_url": "", "post_date": "2023-02-18T00:07:23.963000", "content": "

Interesting!

", "votes": 1, "replies": [] }, { "id": 2148523, "author_name": "KRUPA MISTRY", "author_url": "", "post_date": "2023-02-17T14:10:07.547000", "content": "

Interesting!

", "votes": 1, "replies": [] }, { "id": 2148492, "author_name": "Maria Rosario Lacuna", "author_url": "", "post_date": "2023-02-17T13:28:25.893000", "content": "

Interesting!

", "votes": 1, "replies": [] }, { "id": 2147929, "author_name": "‍김상준[ 학부재학 / 데이터과학과 ]", "author_url": "", "post_date": "2023-02-17T02:50:58.993000", "content": "

This analysis is interesting! It makes me enthusiastic. Thank you.

", "votes": 1, "replies": [] }, { "id": 2147898, "author_name": "Soichiro Murakami", "author_url": "", "post_date": "2023-02-17T01:10:45.573000", "content": "

This insightful analysis helps me a lot! Thank you.

", "votes": 1, "replies": [] }, { "id": 2147623, "author_name": "Konstantin Belousov", "author_url": "", "post_date": "2023-02-16T18:38:54.513000", "content": "

Helpfull thx

", "votes": 1, "replies": [] }, { "id": 2147573, "author_name": "Laurent Pourchot", "author_url": "", "post_date": "2023-02-16T18:01:42.937000", "content": "

@cdeotte , you are amazing….

", "votes": 1, "replies": [] }, { "id": 2146644, "author_name": "Minh T. Nguyen", "author_url": "", "post_date": "2023-02-16T03:08:01.480000", "content": "

@cdeotte awesome notebook and visualization!

", "votes": 1, "replies": [] }, { "id": 2146096, "author_name": "Evdilos_Ikaria", "author_url": "", "post_date": "2023-02-15T15:50:51.250000", "content": "

Thank your Chris for the extra analysis.It seems that this competition is all but a game…

", "votes": 1, "replies": [] }, { "id": 2145734, "author_name": "", "author_url": "", "post_date": "2023-02-15T10:09:38.260000", "content": "", "votes": 1, "replies": [] }, { "id": 2187966, "author_name": "", "author_url": "", "post_date": "2023-03-19T07:33:07.773000", "content": "", "votes": 2, "replies": [] }, { "id": 2186009, "author_name": "", "author_url": "", "post_date": "2023-03-17T13:45:55.050000", "content": "", "votes": 2, "replies": [] }, { "id": 2200230, "author_name": "", "author_url": "", "post_date": "2023-03-28T11:49:56.523000", "content": "", "votes": 0, "replies": [] }, { "id": 2147749, "author_name": "", "author_url": "", "post_date": "2023-02-16T20:59:47.023000", "content": "", "votes": 2, "replies": [ { "id": 2147760, "author_name": "", "author_url": "", "post_date": "2023-02-16T21:22:13.833000", "content": "", "votes": 9, "replies": [] }, { "id": 2147776, "author_name": "", "author_url": "", "post_date": "2023-02-16T21:48:54.887000", "content": "", "votes": 1, "replies": [] }, { "id": 2147780, "author_name": "", "author_url": "", "post_date": "2023-02-16T21:51:26.190000", "content": "", "votes": 3, "replies": [] } ] }, { "id": 2146354, "author_name": "", "author_url": "", "post_date": "2023-02-15T19:59:12.943000", "content": "", "votes": 2, "replies": [] }, { "id": 2145622, "author_name": "", "author_url": "", "post_date": "2023-02-15T08:50:21.647000", "content": "", "votes": 2, "replies": [] }, { "id": 2292271, "author_name": "", "author_url": "", "post_date": "2023-06-08T07:14:41.593000", "content": "", "votes": 0, "replies": [] }, { "id": 2232779, "author_name": "", "author_url": "", "post_date": "2023-04-24T15:50:11.570000", "content": "", "votes": 0, "replies": [] }, { "id": 2219297, "author_name": "", "author_url": "", "post_date": "2023-04-12T13:23:31.143000", "content": "", "votes": 0, "replies": [] }, { "id": 2269416, "author_name": "", "author_url": "", "post_date": "2023-05-22T13:16:11.370000", "content": "", "votes": 0, "replies": [] }, { "id": 2202984, "author_name": "", "author_url": "", "post_date": "2023-03-30T12:45:35.453000", "content": "", "votes": 1, "replies": [] }, { "id": 2202656, "author_name": "", "author_url": "", "post_date": "2023-03-30T08:02:03.773000", "content": "", "votes": 1, "replies": [] }, { "id": 2184740, "author_name": "", "author_url": "", "post_date": "2023-03-16T15:38:03.980000", "content": "", "votes": 1, "replies": [] }, { "id": 2184428, "author_name": "", "author_url": "", "post_date": "2023-03-16T11:29:49.807000", "content": "", "votes": 1, "replies": [] }, { "id": 2183633, "author_name": "", "author_url": "", "post_date": "2023-03-15T19:36:50.257000", "content": "", "votes": 1, "replies": [] }, { "id": 2183055, "author_name": "", "author_url": "", "post_date": "2023-03-15T12:58:26.987000", "content": "", "votes": 1, "replies": [] }, { "id": 2157203, "author_name": "", "author_url": "", "post_date": "2023-02-23T19:26:16.863000", "content": "", "votes": 1, "replies": [] }, { "id": 2156642, "author_name": "", "author_url": "", "post_date": "2023-02-23T13:01:31.690000", "content": "", "votes": 1, "replies": [] }, { "id": 2155857, "author_name": "", "author_url": "", "post_date": "2023-02-23T00:33:27.993000", "content": "", "votes": 1, "replies": [] }, { "id": 2151295, "author_name": "", "author_url": "", "post_date": "2023-02-20T02:04:16.837000", "content": "", "votes": 1, "replies": [] }, { "id": 2149210, "author_name": "", "author_url": "", "post_date": "2023-02-18T04:33:40.670000", "content": "", "votes": 1, "replies": [] }, { "id": 2147709, "author_name": "", "author_url": "", "post_date": "2023-02-16T20:11:18.327000", "content": "", "votes": 1, "replies": [] }, { "id": 2147357, "author_name": "", "author_url": "", "post_date": "2023-02-16T15:03:29.097000", "content": "", "votes": 1, "replies": [] }, { "id": 2147032, "author_name": "", "author_url": "", "post_date": "2023-02-16T11:16:21.163000", "content": "", "votes": 1, "replies": [] }, { "id": 2145499, "author_name": "", "author_url": "", "post_date": "2023-02-15T06:23:06.970000", "content": "", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "2145327": "# How The Game Begins - EDA\nWhen the game begins, our web browser pops out a new window of size 660x880. The following 11 screen shots are the first 11 things we see. We do not have a choice. The game pauses on each image until we click anywhere. \n\nThe first screen we see is the image in the top left (which is also image in yellow box in room tunic.historicalsociety.closet below). No matter where we click, grampa says \"Whatcha doing over there, Jo?\". Then no matter where we click, we say \"Just talking to Teddy.\". The dialog continues as pictured below. (Right click the images below and choose \"Open Image in New Tab\" to enlarge images).\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c12.png)\n\n# First Game Room - tunic.historicalsociety.closet\nEach row of the dataset has 2 pairs of coordinates, `(room_coor_x, room_coor_y)` and `(screen_coor_x, screen_coor_y)`. The first pair is relative to the black rectangle below. And the second pair is relative to your pop out window (not your computer screen). Two examples of pop up window views are the yellow and blue rectanges below. \n\nIn the room tunic.historicalsociety.closet (i.e. the first room), then `-1064<=room_coor_x<=926` and `-393<=room_coor_y<=327`. Each room has its own bounds. The screen coordinates are always bound by your pop out window size. 95% of users do not resize the pop out window and `0<=room_coor_x<=880` and `0<=room_coor_y<=660`\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c2.png)\n\n# First Click Option\nWhen the game begins, we see the image in the yellow rectangle above. No matter where we click, grampa will say \"Whatcha doing over there, Jo?\", but it is still interesting to see where users click first. Some users click `grampa`. Some users click the `retirement_letter` on grampa's desk. See the clusters of points below. Each blue dot is a different users first click:\n### Room Coordinates\nWe plot `plt.scatter(df.room_coor_x, df.room_coor_y)` in the black box:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c3.png)\n### Screen Coordinates\nWe plot `plt.scatter(df.screen_coor_x, 660-df.screen_coor_y)` in the yellow box. Note we flip `screen_coor_y`:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c4.png)\n\n# First Action Option\nAfter we finish the mandatory introductory dialog, we are free to choose our first action. The scatter plot below shows where users make their first action. Most users click Teddy (the lovable pet badger). Some (curious users) click the photos on the wall. Some click to the left of the room to navigate (eager to complete their first quest):\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c5.png)\n\n# First Quest\nThe last message in the dialog (before Teddy's emoji) is Jo saying, \"Now where did I put my notebook?\". This is our first quest. We must find the notebook. If we traverse to the far left side of this first room, we see the notebook on the floor. Every user must find and click this notebook to proceed. Below are every users first click finding the notebook. When we hover over the notebook, we see an orange exclamation point. Some users click the exclamation point, and some users click directly on the notebook:\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c7.png)\n\n# First Use of Notebook\nAfter clicking the notebook, the notebook opens and takes over the full pop up window. The message \"Found it!\" appears in the bottom of the screen over the notebook. To continue the game, we can click anywhere. Most users click on the words \"Found it\", but some users click elsewhere. Regardless of where we click, the message disappears and a white X in black circle appears in top right corner. Most users click the white X in black circle which closes the notebook and play continues. Some users click elsewhere and nothing happens. They must keep clicking until they click the white X in black circle. Some users even click the words \"Homework Monster\" in the bottom center of the notebook!\n\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c8.png)\nWe plot `plt.scatter(df.screen_coor_x, 660-df.screen_coor_y)`. Note we flip `screen_coor_y`:\n![](https://raw.githubusercontent.com/cdeotte/Kaggle_Images/main/Feb-2023/c9.png)\n\n# User Behavior - Feature Engineering\nThere is so much to learn about the 11k users by just observing their behavior in this opening sequence. How long do they take to read each text dialog before continuing? Where do they click each time to continue the game. Where is their first click (grampa, letter, elsewhere)? When users click grampa, do they click grampa's body or face? When Teddy appears on the screen, how many users start clicking the pet versus clicking on Jo to continue the dialog? Which users click text boxes to continue. Which users resize the initial pop up window, which don't? etc etc.\n\nThis is just the opening sequence. There are 19 rooms and 23 levels to explore. Each moment, users take one of 11 actions. There are many dialogs and many quests! Different users click different things. Users reveal much predictive information (i.e. features) by their game playing behavior!\n\n# UPDATE - Starter Notebook\nI published a starter notebook [here][1] displaying click EDA for all 19 game rooms. The starter notebook displays `navigate_clicks` and `person_clicks` and labels items of interest in each room\n\n[1]: https://www.kaggle.com/code/cdeotte/game-room-click-eda", "2145601": "Thank you ! This is informative and well illustrated.\n\nAlso, it's been a long time that I wanted to express that : huge THANK YOU for all your efforts and sharings with the community ! And not only on this competition 🙏", "2145608": "**UPDATE - Starter Notebook**\nI published a starter notebook [here][1] displaying click EDA for all 19 game rooms. The starter notebook displays `navigate_clicks` and `person_clicks` and labels items of interest in each room\n\n[1]: https://www.kaggle.com/cdeotte/game-room-click-eda", "2145526": "I make lots of features based on diff 'elapsed_time', which is really useful. But I have long way to go. Thanks for your impressive illustration.", "2218126": "Very helpful!", "2203674": "Interesting!", "2196424": "It's Brilliant", "2185294": "Very helpful!", "2185254": "What an amazing insight! Just for my curiosity, let me know how did you get it?", "2159116": "Interesting! I actually played the game. It was about the same click.", "2158497": "Nice and helpful !", "2156350": "Interesting", "2155324": "helpful thx ", "2155104": "very helpful！", "2154676": "Thanks for sharing. I hope this will help to understand EDA.", "2154511": "Helpfull thx", "2154377": "very helpful!", "2154087": "Nice notebook", "2153902": "very helpful", "2150307": "Very helpful!", "2149767": "Interesting!", "2149098": "Interesting!", "2148523": "Interesting!", "2148492": "Interesting!", "2147929": "This analysis is interesting! It makes me enthusiastic. Thank you.", "2147898": "This insightful analysis helps me a lot! Thank you.", "2147623": "Helpfull thx", "2147573": "@cdeotte , you are amazing.... ", "2146644": "@cdeotte awesome notebook and visualization!", "2146096": "Thank your Chris for the extra analysis.It seems that this competition is all but a game...", "2145734": "Thanks for sharing this useful information, Chris", "2187966": "Very Interesting!", "2186009": "Nice notebook", "2200230": "Very Interesting!", "2147749": "I found that about half of all sessions have 4 screens different: \"Can I come, Gramps?\" and related 3 more lines of dialogue instead of \"Meetings are BORING!\" and related 3 more lines of of dialogue. There's even some with \"Ugh. Meetings are so boring.\"\n\nBut when I try a session, I always get \"Meetings are BORING!\". If there's a place to click or a secret, I haven't found it yet. EDA didn't show any relevant difference in clicks. There was a big difference in average 'elapsed time'(!) but when I try fast clicking through I didn't get any different result in my personal sessions.\n\n", "2146354": "This is really helpful stuff, Thanks @cdeotte ", "2145622": "Thank you Chris for this awesome analysis! ", "2292271": "Good insights ! Thx!", "2232779": "Your sharing is always so helpful！!", "2219297": "Interesting! @cdeotte\nHow did you know it was that screen? ", "2269416": "", "2202984": "thank you very much", "2202656": "Thanks a lot!👍", "2184740": "Very impressed! Thanks for sharing!", "2184428": "Thanks, it is helpful!", "2183633": "Very informative. Thanks!", "2183055": "Thanks a lot XD", "2157203": "Thanks for your explanation", "2156642": "Thanks for sharing. very helpful!", "2155857": "Thanks for this", "2151295": "very helpful, thanks for sharing!", "2149210": "It's really helpful! Thank you!", "2147709": "very helpful thanks!", "2147357": "Thanks for sharing.", "2147032": "Thanks deotte", "2145499": "Awesome! Thanks a lot Chris." } }