{ "id": 245397, "title": "Are we missing the most important features that drive engagement?", "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/245397", "author_name": "", "post_date": "2021-06-10T20:16:43.262328100Z", "votes": 44, "comment_count": 13, "views": 0, "content": "

Seems like we're missing a lot of the \"features\" that really drive fan engagement nowadays…

\n", "messages": [ { "id": "1344329", "postDate": "06/10/2021 20:16:43", "content": "

Seems like we're missing a lot of the \"features\" that really drive fan engagement nowadays…

\n", "rawMarkdown": "Seems like we're missing a lot of the \"features\" that really drive fan engagement nowadays...\n\n- Did the batter \"pimp\" his home run and/or do an awesome bat flip?\n- Was the pitcher accused of using a foreign substance during or after the game?\n- Was one team caught banging on trash cans?\n- Did someone break one of baseball's many \"unwritten rules\"?\n- Was there a mid-game brawl?\n- Did someone get ejected and go out in an epic fashion like [this](https://www.youtube.com/watch?v=Ggy6WGUFaYs)?", "votes": null }, { "id": "1344795", "postDate": "06/11/2021 06:29:25", "content": "

I could not find data on the coaching staff and refereeing teams

", "rawMarkdown": "I could not find data on the coaching staff and refereeing teams", "votes": null }, { "id": "1346966", "postDate": "06/12/2021 19:22:07", "content": "

@brandenkmurray you are absolutely spot on.

\n

To add…

\n", "rawMarkdown": "brandenkmurray you are absolutely spot on.\n\nTo add...\n- Did they just do a major product endorsement?\n- Did they just have a DUI / arrest?\n- Are they on TMZ for something?", "votes": null }, { "id": "1347200", "postDate": "06/13/2021 04:36:19", "content": "

We have a lot of data on players and teams but it does not seem like the kind of feature(s) that would cause me to grab my phone and twitter/facestuff/tickytokie/whatever that signals engagement.

\n

Sadly for me - I get engaged when a \"head line\" hits my screens as noted by several of the folks here and I wonder what's up????

\n

but I do really hate the idea of needing to put sports news headlines into my model.

", "rawMarkdown": "We have a lot of data on players and teams but it does not seem like the kind of feature(s) that would cause me to grab my phone and twitter/facestuff/tickytokie/whatever that signals engagement.\n\nSadly for me - I get engaged when a \"head line\" hits my screens as noted by several of the folks here and I wonder what's up???? \n\n**but I do really hate the idea of needing to put sports news headlines into my model.**", "votes": null }, { "id": "1349396", "postDate": "06/14/2021 17:54:11", "content": "

While the available datasets for the prediction competition are necessarily constrained to those provided (and defined to be dynamically refreshed for the evaluation period), the explainability prize welcomes you to incorporate any factors for which you can provide analyses that demonstrate their impact on digital engagement. In that analytics competition being offered, feel free to bring in any number of these other datasets!

", "rawMarkdown": "While the available datasets for the prediction competition are necessarily constrained to those provided (and defined to be dynamically refreshed for the evaluation period), the [explainability prize](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/overview/explainability-prize) welcomes you to incorporate any factors for which you can provide analyses that demonstrate their impact on digital engagement. In that analytics competition being offered, feel free to bring in any number of these other datasets!", "votes": null }, { "id": "1350680", "postDate": "06/15/2021 16:39:58", "content": "

It's been 4 days since I joined this still understanding the data.

", "rawMarkdown": "It's been 4 days since I joined this still understanding the data.", "votes": null }, { "id": "1351310", "postDate": "06/16/2021 07:50:29", "content": "

There are many features that could be gained from how much the TV networks are promoting the games. Are the games in the prime time TV slot? Do games get impacted by other TV events during the season? Does TV viewer totals peak at the end of the games as the closing pitches become more important?

\n

This website appears to have the viewer data. Just wondering if there are any historical datasets which can be tested?
\n[https://datarade.ai/data-categories/tv-viewership-data]

", "rawMarkdown": "There are many features that could be gained from how much the TV networks are promoting the games. Are the games in the prime time TV slot? Do games get impacted by other TV events during the season? Does TV viewer totals peak at the end of the games as the closing pitches become more important? \n\nThis website appears to have the viewer data. Just wondering if there are any historical datasets which can be tested?\n[https://datarade.ai/data-categories/tv-viewership-data]", "votes": null }, { "id": "1358456", "postDate": "06/20/2021 13:19:05", "content": "

There are also no factors such as player had an accident or a scandal.
\nI think that if something bad happens to our favorite players, we will be interested.

", "rawMarkdown": "There are also no factors such as player had an accident or a scandal.\nI think that if something bad happens to our favorite players, we will be interested.", "votes": null }, { "id": "1359130", "postDate": "06/21/2021 05:18:43", "content": "

Scraping some known Twitter Baseball accounts' data might be helpful to deal with this problem. As far as I know, using publicly available data is allowed but it might be some work to incorporate it into the main data.

", "rawMarkdown": "Scraping some known Twitter Baseball accounts' data might be helpful to deal with this problem. As far as I know, using publicly available data is allowed but it might be some work to incorporate it into the main data.", "votes": null }, { "id": "1359499", "postDate": "06/21/2021 10:27:07", "content": "

Just checking on this post about \"necessarily constrained to those provided\" and how it relates to the Rules 7C on external data and code requirements \"Freely & publicly available external data is allowed, including pre-trained models\" - does this mean for the code competition only the data provided for the competition may be used? Theoretically external data could be handled in public notebooks and/or public datasets and if they were refreshed, when notebooks are run for evaluation they should pick up the latest versions. Arguably a model in a notebook used as data in the notebooks to be run for evaluation could also be updated.
\nHave not seen anything in the rules or in timeline about how this will be handled, but if no external data is allowed and models need to be created and trained in the submitted notebook only could that be stated in the Rules and discussion thread?

", "rawMarkdown": "Just checking on this post about \"necessarily constrained to those provided\" and how it relates to the Rules 7C on external data and code requirements \"Freely & publicly available external data is allowed, including pre-trained models\" - does this mean for the code competition only the data provided for the competition may be used? Theoretically external data could be handled in public notebooks and/or public datasets and if they were refreshed, when notebooks are run for evaluation they should pick up the latest versions. Arguably a model in a notebook used as data in the notebooks to be run for evaluation could also be updated. \nHave not seen anything in the rules or in timeline about how this will be handled, but if no external data is allowed and models need to be created and trained in the submitted notebook only could that be stated in the Rules and discussion thread?", "votes": null }, { "id": "1366165", "postDate": "06/26/2021 14:08:10", "content": "

So, we can't use external data for the prediction competition? @juliaelliott

", "rawMarkdown": "So, we can't use external data for the prediction competition? @juliaelliott", "votes": null }, { "id": "1366428", "postDate": "06/26/2021 18:44:37", "content": "

RTFB

\n

Rules tab ^

\n
\n

C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).>

\n
", "rawMarkdown": "RTFB\n\nRules tab ^\n\n> C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).>", "votes": null }, { "id": "1368170", "postDate": "06/28/2021 11:26:24", "content": "

@brandenkmurray can we pull in some video to see how many people join a brawl so that would be more @ mentions in tweets? 😄

", "rawMarkdown": "brandenkmurray can we pull in some video to see how many people join a brawl so that would be more @ mentions in tweets? 😄", "votes": null }, { "id": "1369905", "postDate": "06/29/2021 17:35:18", "content": "

@rajeshh In addition to the rules clause shared by PC Jimmmy, I'll add that while training on external data is possible, due to the nature of this code competition, that external data you bring in will not be updated and available for the evaluation period. Only the data indicated on the Data page will be dynamically updated for your code's rerun.

", "rawMarkdown": "rajeshh In addition to the rules clause shared by PC Jimmmy, I'll add that while training on external data is possible, due to the nature of this code competition, that external data you bring in will not be updated and available for the evaluation period. Only the data indicated on the [Data page](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/data) will be dynamically updated for your code's rerun.", "votes": null } ], "comments": [ { "id": 1344795, "author_name": "romankriven", "author_url": "", "post_date": "06/11/2021 06:29:25", "content": "

I could not find data on the coaching staff and refereeing teams

", "votes": null, "replies": [] }, { "id": 1346966, "author_name": "earlred", "author_url": "", "post_date": "06/12/2021 19:22:07", "content": "

@brandenkmurray you are absolutely spot on.

\n

To add…

\n", "votes": null, "replies": [] }, { "id": 1347200, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/13/2021 04:36:19", "content": "

We have a lot of data on players and teams but it does not seem like the kind of feature(s) that would cause me to grab my phone and twitter/facestuff/tickytokie/whatever that signals engagement.

\n

Sadly for me - I get engaged when a \"head line\" hits my screens as noted by several of the folks here and I wonder what's up????

\n

but I do really hate the idea of needing to put sports news headlines into my model.

", "votes": null, "replies": [] }, { "id": 1349396, "author_name": "juliaelliott", "author_url": "", "post_date": "06/14/2021 17:54:11", "content": "

While the available datasets for the prediction competition are necessarily constrained to those provided (and defined to be dynamically refreshed for the evaluation period), the explainability prize welcomes you to incorporate any factors for which you can provide analyses that demonstrate their impact on digital engagement. In that analytics competition being offered, feel free to bring in any number of these other datasets!

", "votes": null, "replies": [ { "id": 1359499, "author_name": "something4kag", "author_url": "", "post_date": "06/21/2021 10:27:07", "content": "

Just checking on this post about \"necessarily constrained to those provided\" and how it relates to the Rules 7C on external data and code requirements \"Freely & publicly available external data is allowed, including pre-trained models\" - does this mean for the code competition only the data provided for the competition may be used? Theoretically external data could be handled in public notebooks and/or public datasets and if they were refreshed, when notebooks are run for evaluation they should pick up the latest versions. Arguably a model in a notebook used as data in the notebooks to be run for evaluation could also be updated.
\nHave not seen anything in the rules or in timeline about how this will be handled, but if no external data is allowed and models need to be created and trained in the submitted notebook only could that be stated in the Rules and discussion thread?

", "votes": null, "replies": [] }, { "id": 1366165, "author_name": "rajeshh", "author_url": "", "post_date": "06/26/2021 14:08:10", "content": "

So, we can't use external data for the prediction competition? @juliaelliott

", "votes": null, "replies": [] }, { "id": 1366428, "author_name": "pcjimmmy", "author_url": "", "post_date": "06/26/2021 18:44:37", "content": "

RTFB

\n

Rules tab ^

\n
\n

C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).>

\n
", "votes": null, "replies": [] }, { "id": 1369905, "author_name": "juliaelliott", "author_url": "", "post_date": "06/29/2021 17:35:18", "content": "

@rajeshh In addition to the rules clause shared by PC Jimmmy, I'll add that while training on external data is possible, due to the nature of this code competition, that external data you bring in will not be updated and available for the evaluation period. Only the data indicated on the Data page will be dynamically updated for your code's rerun.

", "votes": null, "replies": [] } ] }, { "id": 1350680, "author_name": "nelsonjoseph", "author_url": "", "post_date": "06/15/2021 16:39:58", "content": "

It's been 4 days since I joined this still understanding the data.

", "votes": null, "replies": [] }, { "id": 1351310, "author_name": "datajmcn", "author_url": "", "post_date": "06/16/2021 07:50:29", "content": "

There are many features that could be gained from how much the TV networks are promoting the games. Are the games in the prime time TV slot? Do games get impacted by other TV events during the season? Does TV viewer totals peak at the end of the games as the closing pitches become more important?

\n

This website appears to have the viewer data. Just wondering if there are any historical datasets which can be tested?
\n[https://datarade.ai/data-categories/tv-viewership-data]

", "votes": null, "replies": [] }, { "id": 1358456, "author_name": "tsubajiro", "author_url": "", "post_date": "06/20/2021 13:19:05", "content": "

There are also no factors such as player had an accident or a scandal.
\nI think that if something bad happens to our favorite players, we will be interested.

", "votes": null, "replies": [] }, { "id": 1359130, "author_name": "iocak28", "author_url": "", "post_date": "06/21/2021 05:18:43", "content": "

Scraping some known Twitter Baseball accounts' data might be helpful to deal with this problem. As far as I know, using publicly available data is allowed but it might be some work to incorporate it into the main data.

", "votes": null, "replies": [] }, { "id": 1368170, "author_name": "crained", "author_url": "", "post_date": "06/28/2021 11:26:24", "content": "

@brandenkmurray can we pull in some video to see how many people join a brawl so that would be more @ mentions in tweets? 😄

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1344329": "Seems like we're missing a lot of the \"features\" that really drive fan engagement nowadays...\n\n- Did the batter \"pimp\" his home run and/or do an awesome bat flip?\n- Was the pitcher accused of using a foreign substance during or after the game?\n- Was one team caught banging on trash cans?\n- Did someone break one of baseball's many \"unwritten rules\"?\n- Was there a mid-game brawl?\n- Did someone get ejected and go out in an epic fashion like [this](https://www.youtube.com/watch?v=Ggy6WGUFaYs)?", "1344795": "I could not find data on the coaching staff and refereeing teams", "1346966": "brandenkmurray you are absolutely spot on.\n\nTo add...\n- Did they just do a major product endorsement?\n- Did they just have a DUI / arrest?\n- Are they on TMZ for something?", "1347200": "We have a lot of data on players and teams but it does not seem like the kind of feature(s) that would cause me to grab my phone and twitter/facestuff/tickytokie/whatever that signals engagement.\n\nSadly for me - I get engaged when a \"head line\" hits my screens as noted by several of the folks here and I wonder what's up???? \n\n**but I do really hate the idea of needing to put sports news headlines into my model.**", "1349396": "While the available datasets for the prediction competition are necessarily constrained to those provided (and defined to be dynamically refreshed for the evaluation period), the [explainability prize](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/overview/explainability-prize) welcomes you to incorporate any factors for which you can provide analyses that demonstrate their impact on digital engagement. In that analytics competition being offered, feel free to bring in any number of these other datasets!", "1350680": "It's been 4 days since I joined this still understanding the data.", "1351310": "There are many features that could be gained from how much the TV networks are promoting the games. Are the games in the prime time TV slot? Do games get impacted by other TV events during the season? Does TV viewer totals peak at the end of the games as the closing pitches become more important? \n\nThis website appears to have the viewer data. Just wondering if there are any historical datasets which can be tested?\n[https://datarade.ai/data-categories/tv-viewership-data]", "1358456": "There are also no factors such as player had an accident or a scandal.\nI think that if something bad happens to our favorite players, we will be interested.", "1359130": "Scraping some known Twitter Baseball accounts' data might be helpful to deal with this problem. As far as I know, using publicly available data is allowed but it might be some work to incorporate it into the main data.", "1359499": "Just checking on this post about \"necessarily constrained to those provided\" and how it relates to the Rules 7C on external data and code requirements \"Freely & publicly available external data is allowed, including pre-trained models\" - does this mean for the code competition only the data provided for the competition may be used? Theoretically external data could be handled in public notebooks and/or public datasets and if they were refreshed, when notebooks are run for evaluation they should pick up the latest versions. Arguably a model in a notebook used as data in the notebooks to be run for evaluation could also be updated. \nHave not seen anything in the rules or in timeline about how this will be handled, but if no external data is allowed and models need to be created and trained in the submitted notebook only could that be stated in the Rules and discussion thread?", "1366165": "So, we can't use external data for the prediction competition? @juliaelliott", "1366428": "RTFB\n\nRules tab ^\n\n> C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).>", "1368170": "brandenkmurray can we pull in some video to see how many people join a brawl so that would be more @ mentions in tweets? 😄", "1369905": "rajeshh In addition to the rules clause shared by PC Jimmmy, I'll add that while training on external data is possible, due to the nature of this code competition, that external data you bring in will not be updated and available for the evaluation period. Only the data indicated on the [Data page](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting/data) will be dynamically updated for your code's rerun." }, "source": "meta" }