![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F645169%2F8108b9b321ec5be9357090438f95bfaf%2Fbuilding_id-1156-site_id-13-Lodging-residential-electricity-chilledwater-steam.png?generation=1573593723628143&alt=media\")
chilled water , steam, hot water all three influence the consumption of electricity. \nIn fact the dependent variable electricity is dependent on the other three dependent variables.
\n\nKinda model within a model.
\n\nshouldn't we first predict the three variables and use that as an input in the main variable electricity.
\n\nOr am I going wrong somewhere. ( Deviating from path )
", "messages": [ { "id": "670247", "postDate": "11/11/2019 08:02:15", "content": "chilled water , steam, hot water all three influence the consumption of electricity. \nIn fact the dependent variable electricity is dependent on the other three dependent variables.
\n\nKinda model within a model.
\n\nshouldn't we first predict the three variables and use that as an input in the main variable electricity.
\n\nOr am I going wrong somewhere. ( Deviating from path )
", "rawMarkdown": "chilled water , steam, hot water all three influence the consumption of electricity. \nIn fact the dependent variable electricity is dependent on the other three dependent variables.\n\nKinda model within a model.\n\nshouldn't we first predict the three variables and use that as an input in the main variable electricity.\n\nOr am I going wrong somewhere. ( Deviating from path )", "votes": null }, { "id": "670410", "postDate": "11/11/2019 12:40:58", "content": "You're going wrong somewhere. Although the four energy PRODUCTS are not mutually independent, you can't assume they're all produced from the same energy SOURCES. They can be produced from natural gas, fuel oil, coal, etc. Your observation that they interact is correct, however.
", "rawMarkdown": "You're going wrong somewhere. Although the four energy PRODUCTS are not mutually independent, you can't assume they're all produced from the same energy SOURCES. They can be produced from natural gas, fuel oil, coal, etc. Your observation that they interact is correct, however.", "votes": null }, { "id": "670616", "postDate": "11/11/2019 16:50:33", "content": "Thats possible. But not sure anyone will allow producing hot waster through fuel oil or coal. and metering fuel oil and coal seems far fetched. Although Gas is a possibility. Just a hypothetical thought. Not sure it will work in this problem as the test data consists of prediction of all 4 meters
", "rawMarkdown": "Thats possible. But not sure anyone will allow producing hot waster through fuel oil or coal. and metering fuel oil and coal seems far fetched. Although Gas is a possibility. Just a hypothetical thought. Not sure it will work in this problem as the test data consists of prediction of all 4 meters", "votes": null }, { "id": "670681", "postDate": "11/11/2019 18:18:05", "content": "You're making an assumption that the source energy is what's metered. Steam and hot water, no matter what fuel is used, will be metered as the product, i.e. hot water/steam million BTU/hr or KWh. You can't make any assumptions about what fuel source is used. Steam or hot water can be produced by waste heat, electricity, natural gas, fuel oil (diesel), wood waste, etc, but the dataset doesn't reveal that. In the worst case, an electric hot water boiler could be used, and the energy would be counted twice, once as electricity and once as hot water.
", "rawMarkdown": "You're making an assumption that the source energy is what's metered. Steam and hot water, no matter what fuel is used, will be metered as the product, i.e. hot water/steam million BTU/hr or KWh. You can't make any assumptions about what fuel source is used. Steam or hot water can be produced by waste heat, electricity, natural gas, fuel oil (diesel), wood waste, etc, but the dataset doesn't reveal that. In the worst case, an electric hot water boiler could be used, and the energy would be counted twice, once as electricity and once as hot water.", "votes": null }, { "id": "671552", "postDate": "11/12/2019 21:33:18", "content": "In some buildings seems that there is a direct relation between electricity and chilled water:
\n\n
In my notebook the correlation between the different energy aspects is analysed.
", "rawMarkdown": "In some buildings seems that there is a direct relation between electricity and chilled water:\n\n\n\nIn my [notebook](https://www.kaggle.com/juanmah/ashrae-correlation-between-energy-aspects) the correlation between the different energy aspects is analysed.", "votes": null }, { "id": "671690", "postDate": "11/13/2019 03:49:30", "content": "This is cool stuff. Your work brings in lots of insights. Thanks for the same.\n I guess the competition organisers will have to say a word on two on this thought process. As the cause and effect can only be confirmed by them.
\n\nIf chilled water and steam correlates with electricity and we treat the other two as attributes and only electricity as dependent those rows on which electricity is to be metered should influence it.
\n\nHowever considering Larry's thought process all three may rise together due to external factors outside the building which may not be one influencing the other.
\n\nAnother thought is if they are not influencing each other then if one is rising the other should be falling. which is not happening.
\n\nCause and effect is very important
", "rawMarkdown": "This is cool stuff. Your work brings in lots of insights. Thanks for the same.\n I guess the competition organisers will have to say a word on two on this thought process. As the cause and effect can only be confirmed by them. \n\nIf chilled water and steam correlates with electricity and we treat the other two as attributes and only electricity as dependent those rows on which electricity is to be metered should influence it.\n\nHowever considering Larry's thought process all three may rise together due to external factors outside the building which may not be one influencing the other. \n\nAnother thought is if they are not influencing each other then if one is rising the other should be falling. which is not happening.\n\nCause and effect is very important", "votes": null }, { "id": "671710", "postDate": "11/13/2019 04:24:39", "content": "Exactly. Let's look at two possibilities; the buildings are stand-alone or they're part of a district energy system that supplies them chilled water, steam/hot water, and electricity from some centralized plant (many campuses, cities, and large industrial users are supplied this way). If the building is stand-alone, the most likely equipment scenario is cooling with an electric chiller; heating with a steam or hot water boiler, most likely natural gas-fired; and electricity supplied from the local utility. In the heating season, electricity should be base-loaded (for lighting, receptacle loads, etc.) and steam/hot water will be from the boiler (which could be electric. High coincident electricity use could indicate that). If the building is fed from a district energy system, then the chilled water energy will be metered and would be approximately the same as a similar stand-alone building, but the electric use will not show up since it's consumed at the district energy plant. Steam, if supplied from electric central plant boilers would be similar. Same building, entirely different energy profiles. That's why I think that any attempt to generalize energy use on so many undifferentiated buildings is a total waste of time unless more useful significant features are added. The leaked stuff doesn't add that kind of information, BTW.
", "rawMarkdown": "Exactly. Let's look at two possibilities; the buildings are stand-alone or they're part of a district energy system that supplies them chilled water, steam/hot water, and electricity from some centralized plant (many campuses, cities, and large industrial users are supplied this way). If the building is stand-alone, the most likely equipment scenario is cooling with an electric chiller; heating with a steam or hot water boiler, most likely natural gas-fired; and electricity supplied from the local utility. In the heating season, electricity should be base-loaded (for lighting, receptacle loads, etc.) and steam/hot water will be from the boiler (which could be electric. High coincident electricity use could indicate that). If the building is fed from a district energy system, then the chilled water energy will be metered and would be approximately the same as a similar stand-alone building, but the electric use will not show up since it's consumed at the district energy plant. Steam, if supplied from electric central plant boilers would be similar. Same building, entirely different energy profiles. That's why I think that any attempt to generalize energy use on so many undifferentiated buildings is a total waste of time unless more useful significant features are added. The leaked stuff doesn't add that kind of information, BTW.", "votes": null } ], "comments": [ { "id": 670410, "author_name": "ljschuster", "author_url": "", "post_date": "11/11/2019 12:40:58", "content": "You're going wrong somewhere. Although the four energy PRODUCTS are not mutually independent, you can't assume they're all produced from the same energy SOURCES. They can be produced from natural gas, fuel oil, coal, etc. Your observation that they interact is correct, however.
", "votes": null, "replies": [ { "id": 670616, "author_name": "manupandit", "author_url": "", "post_date": "11/11/2019 16:50:33", "content": "Thats possible. But not sure anyone will allow producing hot waster through fuel oil or coal. and metering fuel oil and coal seems far fetched. Although Gas is a possibility. Just a hypothetical thought. Not sure it will work in this problem as the test data consists of prediction of all 4 meters
", "votes": null, "replies": [] } ] }, { "id": 670681, "author_name": "ljschuster", "author_url": "", "post_date": "11/11/2019 18:18:05", "content": "You're making an assumption that the source energy is what's metered. Steam and hot water, no matter what fuel is used, will be metered as the product, i.e. hot water/steam million BTU/hr or KWh. You can't make any assumptions about what fuel source is used. Steam or hot water can be produced by waste heat, electricity, natural gas, fuel oil (diesel), wood waste, etc, but the dataset doesn't reveal that. In the worst case, an electric hot water boiler could be used, and the energy would be counted twice, once as electricity and once as hot water.
", "votes": null, "replies": [] }, { "id": 671552, "author_name": "juanmah", "author_url": "", "post_date": "11/12/2019 21:33:18", "content": "In some buildings seems that there is a direct relation between electricity and chilled water:
\n\n
In my notebook the correlation between the different energy aspects is analysed.
", "votes": null, "replies": [ { "id": 671690, "author_name": "manupandit", "author_url": "", "post_date": "11/13/2019 03:49:30", "content": "This is cool stuff. Your work brings in lots of insights. Thanks for the same.\n I guess the competition organisers will have to say a word on two on this thought process. As the cause and effect can only be confirmed by them.
\n\nIf chilled water and steam correlates with electricity and we treat the other two as attributes and only electricity as dependent those rows on which electricity is to be metered should influence it.
\n\nHowever considering Larry's thought process all three may rise together due to external factors outside the building which may not be one influencing the other.
\n\nAnother thought is if they are not influencing each other then if one is rising the other should be falling. which is not happening.
\n\nCause and effect is very important
", "votes": null, "replies": [] }, { "id": 671710, "author_name": "ljschuster", "author_url": "", "post_date": "11/13/2019 04:24:39", "content": "Exactly. Let's look at two possibilities; the buildings are stand-alone or they're part of a district energy system that supplies them chilled water, steam/hot water, and electricity from some centralized plant (many campuses, cities, and large industrial users are supplied this way). If the building is stand-alone, the most likely equipment scenario is cooling with an electric chiller; heating with a steam or hot water boiler, most likely natural gas-fired; and electricity supplied from the local utility. In the heating season, electricity should be base-loaded (for lighting, receptacle loads, etc.) and steam/hot water will be from the boiler (which could be electric. High coincident electricity use could indicate that). If the building is fed from a district energy system, then the chilled water energy will be metered and would be approximately the same as a similar stand-alone building, but the electric use will not show up since it's consumed at the district energy plant. Steam, if supplied from electric central plant boilers would be similar. Same building, entirely different energy profiles. That's why I think that any attempt to generalize energy use on so many undifferentiated buildings is a total waste of time unless more useful significant features are added. The leaked stuff doesn't add that kind of information, BTW.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "670247": "chilled water , steam, hot water all three influence the consumption of electricity. \nIn fact the dependent variable electricity is dependent on the other three dependent variables.\n\nKinda model within a model.\n\nshouldn't we first predict the three variables and use that as an input in the main variable electricity.\n\nOr am I going wrong somewhere. ( Deviating from path )", "670410": "You're going wrong somewhere. Although the four energy PRODUCTS are not mutually independent, you can't assume they're all produced from the same energy SOURCES. They can be produced from natural gas, fuel oil, coal, etc. Your observation that they interact is correct, however.", "670616": "Thats possible. But not sure anyone will allow producing hot waster through fuel oil or coal. and metering fuel oil and coal seems far fetched. Although Gas is a possibility. Just a hypothetical thought. Not sure it will work in this problem as the test data consists of prediction of all 4 meters", "670681": "You're making an assumption that the source energy is what's metered. Steam and hot water, no matter what fuel is used, will be metered as the product, i.e. hot water/steam million BTU/hr or KWh. You can't make any assumptions about what fuel source is used. Steam or hot water can be produced by waste heat, electricity, natural gas, fuel oil (diesel), wood waste, etc, but the dataset doesn't reveal that. In the worst case, an electric hot water boiler could be used, and the energy would be counted twice, once as electricity and once as hot water.", "671552": "In some buildings seems that there is a direct relation between electricity and chilled water:\n\n\n\nIn my [notebook](https://www.kaggle.com/juanmah/ashrae-correlation-between-energy-aspects) the correlation between the different energy aspects is analysed.", "671690": "This is cool stuff. Your work brings in lots of insights. Thanks for the same.\n I guess the competition organisers will have to say a word on two on this thought process. As the cause and effect can only be confirmed by them. \n\nIf chilled water and steam correlates with electricity and we treat the other two as attributes and only electricity as dependent those rows on which electricity is to be metered should influence it.\n\nHowever considering Larry's thought process all three may rise together due to external factors outside the building which may not be one influencing the other. \n\nAnother thought is if they are not influencing each other then if one is rising the other should be falling. which is not happening.\n\nCause and effect is very important", "671710": "Exactly. Let's look at two possibilities; the buildings are stand-alone or they're part of a district energy system that supplies them chilled water, steam/hot water, and electricity from some centralized plant (many campuses, cities, and large industrial users are supplied this way). If the building is stand-alone, the most likely equipment scenario is cooling with an electric chiller; heating with a steam or hot water boiler, most likely natural gas-fired; and electricity supplied from the local utility. In the heating season, electricity should be base-loaded (for lighting, receptacle loads, etc.) and steam/hot water will be from the boiler (which could be electric. High coincident electricity use could indicate that). If the building is fed from a district energy system, then the chilled water energy will be metered and would be approximately the same as a similar stand-alone building, but the electric use will not show up since it's consumed at the district energy plant. Steam, if supplied from electric central plant boilers would be similar. Same building, entirely different energy profiles. That's why I think that any attempt to generalize energy use on so many undifferentiated buildings is a total waste of time unless more useful significant features are added. The leaked stuff doesn't add that kind of information, BTW." }, "source": "meta" }