![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F473521%2Fcbb2bf212d89fd05dec3d6f249ad33ea%2FScreenshot%202020-02-16%20at%205.08.07%20PM.png?generation=1581840628099694&alt=media\")
I've been looking at the yearly averages for the NO2 from the Sentinel-5P OFFL NO2 dataset and it seems that the power generation facilities are a very small portion of the NO2 being generated.
\n\nThe image below shows the average NO2 across the island, with regions of high NO2 concentration in white. As expected, we see major cities containing a lot of NO2 (car emissions, industrial activities).
\n\n
The next image shows a zoomed in rendering of the main coal generation facility in Puerto Rico (bottom right), responsible for the majority of the island's output. There's a faint white tinge, but compared to the major cities, there is very little NO2 being generated.
\n\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F473521%2F6e67e3ca10a39d04559a058f6181b476%2FScreenshot%202020-02-16%20at%205.08.35%20PM.png?generation=1581840888925590&alt=media\")
If this is really the case, then it's going to be very difficult to get a good estimation for NO2 emissions with this dataset. We won't be able to use any facilities near major cities as the signal would be hidden in all the noise.
\n\nAlso, do we need to care about NO2 emissions of power generation if the overwhelming majority is coming from traffic?
\n\nPlease let me know if you've seen something similar, or if I've made a mistake in my analysis. Thanks!
", "messages": [ { "id": "747304", "postDate": "02/16/2020 08:19:06", "content": "I've been looking at the yearly averages for the NO2 from the Sentinel-5P OFFL NO2 dataset and it seems that the power generation facilities are a very small portion of the NO2 being generated.
\n\nThe image below shows the average NO2 across the island, with regions of high NO2 concentration in white. As expected, we see major cities containing a lot of NO2 (car emissions, industrial activities).
\n\n
The next image shows a zoomed in rendering of the main coal generation facility in Puerto Rico (bottom right), responsible for the majority of the island's output. There's a faint white tinge, but compared to the major cities, there is very little NO2 being generated.
\n\n
If this is really the case, then it's going to be very difficult to get a good estimation for NO2 emissions with this dataset. We won't be able to use any facilities near major cities as the signal would be hidden in all the noise.
\n\nAlso, do we need to care about NO2 emissions of power generation if the overwhelming majority is coming from traffic?
\n\nPlease let me know if you've seen something similar, or if I've made a mistake in my analysis. Thanks!
", "rawMarkdown": "I've been looking at the yearly averages for the NO2 from the Sentinel-5P OFFL NO2 dataset and it seems that the power generation facilities are a very small portion of the NO2 being generated.\n\nThe image below shows the average NO2 across the island, with regions of high NO2 concentration in white. As expected, we see major cities containing a lot of NO2 (car emissions, industrial activities).\n\n\n\nThe next image shows a zoomed in rendering of the main coal generation facility in Puerto Rico (bottom right), responsible for the majority of the island's output. There's a faint white tinge, but compared to the major cities, there is very little NO2 being generated.\n\n\n\nIf this is really the case, then it's going to be very difficult to get a good estimation for NO2 emissions with this dataset. We won't be able to use any facilities near major cities as the signal would be hidden in all the noise.\n\nAlso, do we need to care about NO2 emissions of power generation if the overwhelming majority is coming from traffic?\n\nPlease let me know if you've seen something similar, or if I've made a mistake in my analysis. Thanks!", "votes": null }, { "id": "747567", "postDate": "02/16/2020 15:18:31", "content": "The annual emissions report for the USA is in draft form right now for 2018. If you read all 700+ pages you might walk away with the impression that a ton of folks spend a lot of time making guesses that end up being the current reported values.
\n\nThe list of things that put emissions into the air is huge - your correct that power plants are not the hugest slice of the pie - but they are the largest slice where a remote sensing (satellite data) has some reasonable chance of being modeled. They sit in a fixed location, the current estimates of there emissions are probably half way accurate and make a good starting point for perhaps changing thousands of hours of spread sheet cranking to build the annual USA report two years after the fact to a report that might be generated this week for last months performance.
", "rawMarkdown": "The annual emissions report for the USA is in draft form right now for 2018. If you read all 700+ pages you might walk away with the impression that a ton of folks spend a lot of time making guesses that end up being the current reported values. \n\nThe list of things that put emissions into the air is huge - your correct that power plants are not the hugest slice of the pie - but they are the largest slice where a remote sensing (satellite data) has some reasonable chance of being modeled. They sit in a fixed location, the current estimates of there emissions are probably half way accurate and make a good starting point for perhaps changing thousands of hours of spread sheet cranking to build the annual USA report two years after the fact to a report that might be generated this week for last months performance.", "votes": null }, { "id": "747690", "postDate": "02/16/2020 18:14:24", "content": "Focusing on the large power plants on the south of the island (that are far away from the population centers) is one way to isolate the signal from the noise. These are great things to describe in your submission -- what challenges you encountered and how you decided what was the best approach to solve them.
\n\nYour screenshots are very helpful to illustrate that the city of San Juan seems to emit the majority of the NO2 in Puerto Rico. Do you think that is mostly from vehicles or maybe small industry? A supplementary dataset that describes the percent of NO2 emissions from electricity production compared to other industries would be an informative dataset to attach to your submission.
", "rawMarkdown": "Focusing on the large power plants on the south of the island (that are far away from the population centers) is one way to isolate the signal from the noise. These are great things to describe in your submission -- what challenges you encountered and how you decided what was the best approach to solve them. \n\nYour screenshots are very helpful to illustrate that the city of San Juan seems to emit the majority of the NO2 in Puerto Rico. Do you think that is mostly from vehicles or maybe small industry? A supplementary dataset that describes the percent of NO2 emissions from electricity production compared to other industries would be an informative dataset to attach to your submission.", "votes": null }, { "id": "753285", "postDate": "02/22/2020 01:12:34", "content": "Thanks for the detailed response, it's interesting to understand a bit more about why the competition exists.
\n\nHaving studied civil engineering I can relate to making 'guesses' for things that end up being the reported values. Hopefully through the competition we can develop some improved methods based on this awesome satellite data.
", "rawMarkdown": "Thanks for the detailed response, it's interesting to understand a bit more about why the competition exists.\n\nHaving studied civil engineering I can relate to making 'guesses' for things that end up being the reported values. Hopefully through the competition we can develop some improved methods based on this awesome satellite data.", "votes": null }, { "id": "753287", "postDate": "02/22/2020 01:16:25", "content": "Thanks, I think there are many things that can be attempted, e.g. isolating the different types of emissions that are more strongly associated with different sources to remove baselines or looking more into the temporal features such as time of day or day of week of the reading where the emissions for different sources could be at different strengths. Time to dig!
", "rawMarkdown": "Thanks, I think there are many things that can be attempted, e.g. isolating the different types of emissions that are more strongly associated with different sources to remove baselines or looking more into the temporal features such as time of day or day of week of the reading where the emissions for different sources could be at different strengths. Time to dig!", "votes": null } ], "comments": [ { "id": 747567, "author_name": "pcjimmmy", "author_url": "", "post_date": "02/16/2020 15:18:31", "content": "The annual emissions report for the USA is in draft form right now for 2018. If you read all 700+ pages you might walk away with the impression that a ton of folks spend a lot of time making guesses that end up being the current reported values.
\n\nThe list of things that put emissions into the air is huge - your correct that power plants are not the hugest slice of the pie - but they are the largest slice where a remote sensing (satellite data) has some reasonable chance of being modeled. They sit in a fixed location, the current estimates of there emissions are probably half way accurate and make a good starting point for perhaps changing thousands of hours of spread sheet cranking to build the annual USA report two years after the fact to a report that might be generated this week for last months performance.
", "votes": null, "replies": [ { "id": 753285, "author_name": "jyesawtellrickson", "author_url": "", "post_date": "02/22/2020 01:12:34", "content": "Thanks for the detailed response, it's interesting to understand a bit more about why the competition exists.
\n\nHaving studied civil engineering I can relate to making 'guesses' for things that end up being the reported values. Hopefully through the competition we can develop some improved methods based on this awesome satellite data.
", "votes": null, "replies": [] } ] }, { "id": 747690, "author_name": "paultimothymooney", "author_url": "", "post_date": "02/16/2020 18:14:24", "content": "Focusing on the large power plants on the south of the island (that are far away from the population centers) is one way to isolate the signal from the noise. These are great things to describe in your submission -- what challenges you encountered and how you decided what was the best approach to solve them.
\n\nYour screenshots are very helpful to illustrate that the city of San Juan seems to emit the majority of the NO2 in Puerto Rico. Do you think that is mostly from vehicles or maybe small industry? A supplementary dataset that describes the percent of NO2 emissions from electricity production compared to other industries would be an informative dataset to attach to your submission.
", "votes": null, "replies": [ { "id": 753287, "author_name": "jyesawtellrickson", "author_url": "", "post_date": "02/22/2020 01:16:25", "content": "Thanks, I think there are many things that can be attempted, e.g. isolating the different types of emissions that are more strongly associated with different sources to remove baselines or looking more into the temporal features such as time of day or day of week of the reading where the emissions for different sources could be at different strengths. Time to dig!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "747304": "I've been looking at the yearly averages for the NO2 from the Sentinel-5P OFFL NO2 dataset and it seems that the power generation facilities are a very small portion of the NO2 being generated.\n\nThe image below shows the average NO2 across the island, with regions of high NO2 concentration in white. As expected, we see major cities containing a lot of NO2 (car emissions, industrial activities).\n\n\n\nThe next image shows a zoomed in rendering of the main coal generation facility in Puerto Rico (bottom right), responsible for the majority of the island's output. There's a faint white tinge, but compared to the major cities, there is very little NO2 being generated.\n\n\n\nIf this is really the case, then it's going to be very difficult to get a good estimation for NO2 emissions with this dataset. We won't be able to use any facilities near major cities as the signal would be hidden in all the noise.\n\nAlso, do we need to care about NO2 emissions of power generation if the overwhelming majority is coming from traffic?\n\nPlease let me know if you've seen something similar, or if I've made a mistake in my analysis. Thanks!", "747567": "The annual emissions report for the USA is in draft form right now for 2018. If you read all 700+ pages you might walk away with the impression that a ton of folks spend a lot of time making guesses that end up being the current reported values. \n\nThe list of things that put emissions into the air is huge - your correct that power plants are not the hugest slice of the pie - but they are the largest slice where a remote sensing (satellite data) has some reasonable chance of being modeled. They sit in a fixed location, the current estimates of there emissions are probably half way accurate and make a good starting point for perhaps changing thousands of hours of spread sheet cranking to build the annual USA report two years after the fact to a report that might be generated this week for last months performance.", "747690": "Focusing on the large power plants on the south of the island (that are far away from the population centers) is one way to isolate the signal from the noise. These are great things to describe in your submission -- what challenges you encountered and how you decided what was the best approach to solve them. \n\nYour screenshots are very helpful to illustrate that the city of San Juan seems to emit the majority of the NO2 in Puerto Rico. Do you think that is mostly from vehicles or maybe small industry? A supplementary dataset that describes the percent of NO2 emissions from electricity production compared to other industries would be an informative dataset to attach to your submission.", "753285": "Thanks for the detailed response, it's interesting to understand a bit more about why the competition exists.\n\nHaving studied civil engineering I can relate to making 'guesses' for things that end up being the reported values. Hopefully through the competition we can develop some improved methods based on this awesome satellite data.", "753287": "Thanks, I think there are many things that can be attempted, e.g. isolating the different types of emissions that are more strongly associated with different sources to remove baselines or looking more into the temporal features such as time of day or day of week of the reading where the emissions for different sources could be at different strengths. Time to dig!" }, "source": "meta" }