I noticed the rules state that only the given data can be used unless specifically told otherwise. This seems kind of limiting and I've already noticed a few people incorporating some additional fields from posts on the forums. My question is, What external data can we use? Can we tag customers with regions or other city/town related data? Can we add product categories? I think I noticed one post where some one was adding brand, is this allowed?
", "messages": [ { "id": "124271", "postDate": "06/16/2016 17:05:31", "content": "I noticed the rules state that only the given data can be used unless specifically told otherwise. This seems kind of limiting and I've already noticed a few people incorporating some additional fields from posts on the forums. My question is, What external data can we use? Can we tag customers with regions or other city/town related data? Can we add product categories? I think I noticed one post where some one was adding brand, is this allowed?
", "rawMarkdown": "I noticed the rules state that only the given data can be used unless specifically told otherwise. This seems kind of limiting and I've already noticed a few people incorporating some additional fields from posts on the forums. My question is, What external data can we use? Can we tag customers with regions or other city/town related data? Can we add product categories? I think I noticed one post where some one was adding brand, is this allowed?", "votes": null }, { "id": "124289", "postDate": "06/16/2016 19:56:56", "content": "Everything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.
", "rawMarkdown": "Everything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.", "votes": null }, { "id": "124290", "postDate": "06/16/2016 20:27:28", "content": "I wonder what was the reason to prohibit external data, and also what was the reason to hide the real dates? Is it for regulatory compliance or to avoid disclosing trade secrets or both?
", "rawMarkdown": "I wonder what was the reason to prohibit external data, and also what was the reason to hide the real dates? Is it for regulatory compliance or to avoid disclosing trade secrets or both?", "votes": null }, { "id": "124293", "postDate": "06/16/2016 20:52:04", "content": "External data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.
", "rawMarkdown": "External data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.", "votes": null }, { "id": "124356", "postDate": "06/17/2016 13:40:49", "content": "[quote=kes367;124289]
\n\nEverything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.
\n\n[/quote]
\n\nWell that's where I'm not too sure. It may be based off the spreadsheets but we'll certainly have to use other sources to create those factors. Product categories would likely be fine, because they are somewhat subjective. Adding a region code, holiday flag, or brand would involve external research. And adding any type of city data, such as income or population would almost certainly be external.
", "rawMarkdown": "[quote=kes367;124289]\r\n\r\nEverything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.\r\n\r\n[/quote]\r\n\r\nWell that's where I'm not too sure. It may be based off the spreadsheets but we'll certainly have to use other sources to create those factors. Product categories would likely be fine, because they are somewhat subjective. Adding a region code, holiday flag, or brand would involve external research. And adding any type of city data, such as income or population would almost certainly be external.", "votes": null }, { "id": "124357", "postDate": "06/17/2016 13:45:18", "content": "The brand is from the spreadsheets provided. Each line of text has the brand name, product name, weight, pieces, etc in it (in the producto_tabla.csv). What that script creator did was parse through each line, separate each of these out, and organized it into a nice mergeable table. If it can't be directly derived from the spreadsheets given by Grupo Bimbo, then it's likely external information, and thus prohibited.
", "rawMarkdown": "The brand is from the spreadsheets provided. Each line of text has the brand name, product name, weight, pieces, etc in it (in the producto_tabla.csv). What that script creator did was parse through each line, separate each of these out, and organized it into a nice mergeable table. If it can't be directly derived from the spreadsheets given by Grupo Bimbo, then it's likely external information, and thus prohibited.", "votes": null }, { "id": "124358", "postDate": "06/17/2016 13:46:36", "content": "[quote=Branden Murray;124293]
\n\nExternal data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.
\n\n[/quote]
\n\nYou make a good point about preventing the use of other published sales data. My main question is can I use city information, such as demographics and such(along with brand). I figured it would be a question for the Kaggle or Grupo Bimbo, as this is my first competition where i'm doing more than just looking I'm not too sure the best place to ask them questions. Would it be this forum?
", "rawMarkdown": "[quote=Branden Murray;124293]\r\n\r\nExternal data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.\r\n\r\n[/quote]\r\n\r\nYou make a good point about preventing the use of other published sales data. My main question is can I use city information, such as demographics and such(along with brand). I figured it would be a question for the Kaggle or Grupo Bimbo, as this is my first competition where i'm doing more than just looking I'm not too sure the best place to ask them questions. Would it be this forum?", "votes": null }, { "id": "124359", "postDate": "06/17/2016 14:23:52", "content": "Unless it is explicitly and specifically mentioned for this competition, you cannot use any form of external data, publicly available or not.
\n\nSo, no, city information and demographics would not be allowed for this competition.
\n\nAnd, yes, this is probably the best place to ask questions :)
", "rawMarkdown": "Unless it is explicitly and specifically mentioned for this competition, you cannot use any form of external data, publicly available or not. \r\n\r\nSo, no, city information and demographics would not be allowed for this competition. \r\n\r\nAnd, yes, this is probably the best place to ask questions :)", "votes": null } ], "comments": [ { "id": 124289, "author_name": "kes367", "author_url": "", "post_date": "06/16/2016 19:56:56", "content": "Everything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.
", "votes": null, "replies": [] }, { "id": 124290, "author_name": "vykhand", "author_url": "", "post_date": "06/16/2016 20:27:28", "content": "I wonder what was the reason to prohibit external data, and also what was the reason to hide the real dates? Is it for regulatory compliance or to avoid disclosing trade secrets or both?
", "votes": null, "replies": [] }, { "id": 124293, "author_name": "brandenkmurray", "author_url": "", "post_date": "06/16/2016 20:52:04", "content": "External data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.
", "votes": null, "replies": [] }, { "id": 124356, "author_name": "kyle7day", "author_url": "", "post_date": "06/17/2016 13:40:49", "content": "[quote=kes367;124289]
\n\nEverything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.
\n\n[/quote]
\n\nWell that's where I'm not too sure. It may be based off the spreadsheets but we'll certainly have to use other sources to create those factors. Product categories would likely be fine, because they are somewhat subjective. Adding a region code, holiday flag, or brand would involve external research. And adding any type of city data, such as income or population would almost certainly be external.
", "votes": null, "replies": [] }, { "id": 124357, "author_name": "kes367", "author_url": "", "post_date": "06/17/2016 13:45:18", "content": "The brand is from the spreadsheets provided. Each line of text has the brand name, product name, weight, pieces, etc in it (in the producto_tabla.csv). What that script creator did was parse through each line, separate each of these out, and organized it into a nice mergeable table. If it can't be directly derived from the spreadsheets given by Grupo Bimbo, then it's likely external information, and thus prohibited.
", "votes": null, "replies": [] }, { "id": 124358, "author_name": "kyle7day", "author_url": "", "post_date": "06/17/2016 13:46:36", "content": "[quote=Branden Murray;124293]
\n\nExternal data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.
\n\n[/quote]
\n\nYou make a good point about preventing the use of other published sales data. My main question is can I use city information, such as demographics and such(along with brand). I figured it would be a question for the Kaggle or Grupo Bimbo, as this is my first competition where i'm doing more than just looking I'm not too sure the best place to ask them questions. Would it be this forum?
", "votes": null, "replies": [] }, { "id": 124359, "author_name": "gpistre", "author_url": "", "post_date": "06/17/2016 14:23:52", "content": "Unless it is explicitly and specifically mentioned for this competition, you cannot use any form of external data, publicly available or not.
\n\nSo, no, city information and demographics would not be allowed for this competition.
\n\nAnd, yes, this is probably the best place to ask questions :)
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "124271": "I noticed the rules state that only the given data can be used unless specifically told otherwise. This seems kind of limiting and I've already noticed a few people incorporating some additional fields from posts on the forums. My question is, What external data can we use? Can we tag customers with regions or other city/town related data? Can we add product categories? I think I noticed one post where some one was adding brand, is this allowed?", "124289": "Everything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.", "124290": "I wonder what was the reason to prohibit external data, and also what was the reason to hide the real dates? Is it for regulatory compliance or to avoid disclosing trade secrets or both?", "124293": "External data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.", "124356": "[quote=kes367;124289]\r\n\r\nEverything that you mentioned has been derived from the spreadsheets that Grupo Bimbo has provided, it appears. So none of it is external data.\r\n\r\n[/quote]\r\n\r\nWell that's where I'm not too sure. It may be based off the spreadsheets but we'll certainly have to use other sources to create those factors. Product categories would likely be fine, because they are somewhat subjective. Adding a region code, holiday flag, or brand would involve external research. And adding any type of city data, such as income or population would almost certainly be external.", "124357": "The brand is from the spreadsheets provided. Each line of text has the brand name, product name, weight, pieces, etc in it (in the producto_tabla.csv). What that script creator did was parse through each line, separate each of these out, and organized it into a nice mergeable table. If it can't be directly derived from the spreadsheets given by Grupo Bimbo, then it's likely external information, and thus prohibited.", "124358": "[quote=Branden Murray;124293]\r\n\r\nExternal data is prohibited from most competitions. Only Kaggle and Grupo Bimbo can answer why it's the case for this competition but my guess is that it's generally prohibited A) to keep a level playing field, B) it forces us to extract as much as we possibly can out of the data that's currently available to the sponsor, and C) many solutions are already very complex so adding external data to it makes assessing winner's solutions more difficult. With C), for example, for this competition if we knew the dates then we could potentially find bread sales data for different areas of Mexico, and we'd have that data for the test weeks 10 and 11, even though in reality we wouldn't have such data at the time of forecasting. Solutions utilizing that data should probably be disqualified, but if there's a bunch of different external data sets being used it might be hard to see that someone is using that in their solution.\r\n\r\n[/quote]\r\n\r\nYou make a good point about preventing the use of other published sales data. My main question is can I use city information, such as demographics and such(along with brand). I figured it would be a question for the Kaggle or Grupo Bimbo, as this is my first competition where i'm doing more than just looking I'm not too sure the best place to ask them questions. Would it be this forum?", "124359": "Unless it is explicitly and specifically mentioned for this competition, you cannot use any form of external data, publicly available or not. \r\n\r\nSo, no, city information and demographics would not be allowed for this competition. \r\n\r\nAnd, yes, this is probably the best place to ask questions :)" }, "source": "meta" }