I imagined the check-in hubs (place_id) when plotted would look something like a circle with high accuracy points near to the center and lower accuracy points far away from the center.
\n\nBut after actually plotting few check-in hubs, it looks like FB have give us the data of passengers checking in railway station at different platforms.
\n\nThanks to msjgriffiths for sharing the lovely data exploration.
", "messages": [ { "id": "119866", "postDate": "05/13/2016 10:38:52", "content": "I imagined the check-in hubs (place_id) when plotted would look something like a circle with high accuracy points near to the center and lower accuracy points far away from the center.
\n\nBut after actually plotting few check-in hubs, it looks like FB have give us the data of passengers checking in railway station at different platforms.
\n\nThanks to msjgriffiths for sharing the lovely data exploration.
", "rawMarkdown": "I imagined the check-in hubs (place_id) when plotted would look something like a circle with high accuracy points near to the center and lower accuracy points far away from the center.\r\n\r\nBut after actually plotting few check-in hubs, it looks like FB have give us the data of passengers checking in railway station at different platforms. \r\n\r\nThanks to msjgriffiths for sharing the lovely data exploration.", "votes": null }, { "id": "119885", "postDate": "05/13/2016 15:03:25", "content": "@Brox I like it, it fits. I was thinking the orientation of roads, but given that most towns have roads that are perpendicular ..... yeah, you're probably right.!
\n\nPeople had pointed out that var(x) > var(y) for a given place ID, I think (check some of the scripts). It's also somewhat evident from the data description page (in the image, you can kinda see it, but not as clearly in your image).
\n\nSomeone had hypothesized that the imagined city could be San Fran. Does your hypothesis fit with the layout of public transit in that area?
", "rawMarkdown": "Brox I like it, it fits. I was thinking the orientation of roads, but given that most towns have roads that are perpendicular ..... yeah, you're probably right.!\r\n\r\nPeople had pointed out that var(x) > var(y) for a given place ID, I think (check some of the scripts). It's also somewhat evident from the data description page (in the image, you can kinda see it, but not as clearly in your image).\r\n\r\nSomeone had hypothesized that the imagined city could be San Fran. Does your hypothesis fit with the layout of public transit in that area?", "votes": null }, { "id": "119890", "postDate": "05/13/2016 15:22:17", "content": "train tracks across 10 km? big ass platform :P
", "rawMarkdown": "train tracks across 10 km? big ass platform :P", "votes": null }, { "id": "119897", "postDate": "05/13/2016 15:44:28", "content": "I wasn't thinking it would all be on the same platform. But maybe in an area with dense stations. That said, in such a case you wouldn't expect the tracks to be straight for 10 km, would you? Also, to get the variation in the y-direction (not the direction of the tracks), you would need to be at a platform, like you said.
\n\nBack to thinking about roads.
\n\nhttps://www.google.com/maps/@40.7474276,-73.9803644,15.9z
\n\nLook at Manhattan. The blocks are rectangular. A lot, but not all, of the buildings are also rectangular. Similar looks to be true for San Fran. I'm an ecologist, this feature of urban areas wasn't obvious to me, ha.
", "rawMarkdown": "I wasn't thinking it would all be on the same platform. But maybe in an area with dense stations. That said, in such a case you wouldn't expect the tracks to be straight for 10 km, would you? Also, to get the variation in the y-direction (not the direction of the tracks), you would need to be at a platform, like you said.\r\n\r\nBack to thinking about roads.\r\n\r\nhttps://www.google.com/maps/@40.7474276,-73.9803644,15.9z\r\n\r\nLook at Manhattan. The blocks are rectangular. A lot, but not all, of the buildings are also rectangular. Similar looks to be true for San Fran. I'm an ecologist, this feature of urban areas wasn't obvious to me, ha.", "votes": null }, { "id": "119976", "postDate": "05/14/2016 08:13:01", "content": "My theory is there are say 5 stations say each 2 kms apart. Some check in as soon as they reach the platform, some check in after they get in to the train and some check in when they are in transit.
\n\nso it's quite possible that there are checkins all over the railway line.
", "rawMarkdown": "My theory is there are say 5 stations say each 2 kms apart. Some check in as soon as they reach the platform, some check in after they get in to the train and some check in when they are in transit. \r\n\r\nso it's quite possible that there are checkins all over the railway line.", "votes": null }, { "id": "119978", "postDate": "05/14/2016 08:29:52", "content": "Railway platforms or otherwise, it is safe to assume that place_id's are not to be treated as point objects. Thanks for the observation.
", "rawMarkdown": "Railway platforms or otherwise, it is safe to assume that place_id's are not to be treated as point objects. Thanks for the observation.", "votes": null }, { "id": "120476", "postDate": "05/18/2016 14:54:12", "content": "I'm glad you posted that picture. I was about to post something that looked the same. I was starting to think I couldn't even read in a simple data file.
\n\nAttached is a plot of 10 random place_ids.
", "rawMarkdown": "I'm glad you posted that picture. I was about to post something that looked the same. I was starting to think I couldn't even read in a simple data file.\r\n\r\nAttached is a plot of 10 random place_ids.", "votes": null }, { "id": "120494", "postDate": "05/18/2016 16:46:00", "content": "I don't really think it's something related with the shape of the places. Simply because the large variation is (almost?) always in the X axis, and because it is affecting a large number of the places. So, unless we have in front of us an artificial world of streets/railway platforms/whatever all oriented from west to east (and none on from north to south), my bet would be that we have something else...
\n\nNow that we know that timestamps are in minutes, maybe it would be worth to see how variations in the X axis correlate with the time of the day/minute of the hour/day of the week, to see if there's some pattern (from other visualisations it is somewhat clear that there's no correlation with the absolute time value). Maybe...
", "rawMarkdown": "I don't really think it's something related with the shape of the places. Simply because the large variation is (almost?) always in the X axis, and because it is affecting a large number of the places. So, unless we have in front of us an artificial world of streets/railway platforms/whatever all oriented from west to east (and none on from north to south), my bet would be that we have something else...\r\n\r\nNow that we know that timestamps are in minutes, maybe it would be worth to see how variations in the X axis correlate with the time of the day/minute of the hour/day of the week, to see if there's some pattern (from other visualisations it is somewhat clear that there's no correlation with the absolute time value). Maybe...", "votes": null }, { "id": "121539", "postDate": "05/27/2016 05:21:35", "content": "@carrDelling yes, now that we have some information from "minutes" we should be able to better guess what these places are.
", "rawMarkdown": "carrDelling yes, now that we have some information from \"minutes\" we should be able to better guess what these places are.", "votes": null } ], "comments": [ { "id": 119885, "author_name": "zeraes", "author_url": "", "post_date": "05/13/2016 15:03:25", "content": "@Brox I like it, it fits. I was thinking the orientation of roads, but given that most towns have roads that are perpendicular ..... yeah, you're probably right.!
\n\nPeople had pointed out that var(x) > var(y) for a given place ID, I think (check some of the scripts). It's also somewhat evident from the data description page (in the image, you can kinda see it, but not as clearly in your image).
\n\nSomeone had hypothesized that the imagined city could be San Fran. Does your hypothesis fit with the layout of public transit in that area?
", "votes": null, "replies": [] }, { "id": 119890, "author_name": "", "author_url": "", "post_date": "05/13/2016 15:22:17", "content": "train tracks across 10 km? big ass platform :P
", "votes": null, "replies": [] }, { "id": 119897, "author_name": "zeraes", "author_url": "", "post_date": "05/13/2016 15:44:28", "content": "I wasn't thinking it would all be on the same platform. But maybe in an area with dense stations. That said, in such a case you wouldn't expect the tracks to be straight for 10 km, would you? Also, to get the variation in the y-direction (not the direction of the tracks), you would need to be at a platform, like you said.
\n\nBack to thinking about roads.
\n\nhttps://www.google.com/maps/@40.7474276,-73.9803644,15.9z
\n\nLook at Manhattan. The blocks are rectangular. A lot, but not all, of the buildings are also rectangular. Similar looks to be true for San Fran. I'm an ecologist, this feature of urban areas wasn't obvious to me, ha.
", "votes": null, "replies": [] }, { "id": 119976, "author_name": "", "author_url": "", "post_date": "05/14/2016 08:13:01", "content": "My theory is there are say 5 stations say each 2 kms apart. Some check in as soon as they reach the platform, some check in after they get in to the train and some check in when they are in transit.
\n\nso it's quite possible that there are checkins all over the railway line.
", "votes": null, "replies": [] }, { "id": 119978, "author_name": "adarshchavakula", "author_url": "", "post_date": "05/14/2016 08:29:52", "content": "Railway platforms or otherwise, it is safe to assume that place_id's are not to be treated as point objects. Thanks for the observation.
", "votes": null, "replies": [] }, { "id": 120476, "author_name": "muenzer", "author_url": "", "post_date": "05/18/2016 14:54:12", "content": "I'm glad you posted that picture. I was about to post something that looked the same. I was starting to think I couldn't even read in a simple data file.
\n\nAttached is a plot of 10 random place_ids.
", "votes": null, "replies": [] }, { "id": 120494, "author_name": "carrdelling", "author_url": "", "post_date": "05/18/2016 16:46:00", "content": "I don't really think it's something related with the shape of the places. Simply because the large variation is (almost?) always in the X axis, and because it is affecting a large number of the places. So, unless we have in front of us an artificial world of streets/railway platforms/whatever all oriented from west to east (and none on from north to south), my bet would be that we have something else...
\n\nNow that we know that timestamps are in minutes, maybe it would be worth to see how variations in the X axis correlate with the time of the day/minute of the hour/day of the week, to see if there's some pattern (from other visualisations it is somewhat clear that there's no correlation with the absolute time value). Maybe...
", "votes": null, "replies": [] }, { "id": 121539, "author_name": "", "author_url": "", "post_date": "05/27/2016 05:21:35", "content": "@carrDelling yes, now that we have some information from "minutes" we should be able to better guess what these places are.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "119866": "I imagined the check-in hubs (place_id) when plotted would look something like a circle with high accuracy points near to the center and lower accuracy points far away from the center.\r\n\r\nBut after actually plotting few check-in hubs, it looks like FB have give us the data of passengers checking in railway station at different platforms. \r\n\r\nThanks to msjgriffiths for sharing the lovely data exploration.", "119885": "Brox I like it, it fits. I was thinking the orientation of roads, but given that most towns have roads that are perpendicular ..... yeah, you're probably right.!\r\n\r\nPeople had pointed out that var(x) > var(y) for a given place ID, I think (check some of the scripts). It's also somewhat evident from the data description page (in the image, you can kinda see it, but not as clearly in your image).\r\n\r\nSomeone had hypothesized that the imagined city could be San Fran. Does your hypothesis fit with the layout of public transit in that area?", "119890": "train tracks across 10 km? big ass platform :P", "119897": "I wasn't thinking it would all be on the same platform. But maybe in an area with dense stations. That said, in such a case you wouldn't expect the tracks to be straight for 10 km, would you? Also, to get the variation in the y-direction (not the direction of the tracks), you would need to be at a platform, like you said.\r\n\r\nBack to thinking about roads.\r\n\r\nhttps://www.google.com/maps/@40.7474276,-73.9803644,15.9z\r\n\r\nLook at Manhattan. The blocks are rectangular. A lot, but not all, of the buildings are also rectangular. Similar looks to be true for San Fran. I'm an ecologist, this feature of urban areas wasn't obvious to me, ha.", "119976": "My theory is there are say 5 stations say each 2 kms apart. Some check in as soon as they reach the platform, some check in after they get in to the train and some check in when they are in transit. \r\n\r\nso it's quite possible that there are checkins all over the railway line.", "119978": "Railway platforms or otherwise, it is safe to assume that place_id's are not to be treated as point objects. Thanks for the observation.", "120476": "I'm glad you posted that picture. I was about to post something that looked the same. I was starting to think I couldn't even read in a simple data file.\r\n\r\nAttached is a plot of 10 random place_ids.", "120494": "I don't really think it's something related with the shape of the places. Simply because the large variation is (almost?) always in the X axis, and because it is affecting a large number of the places. So, unless we have in front of us an artificial world of streets/railway platforms/whatever all oriented from west to east (and none on from north to south), my bet would be that we have something else...\r\n\r\nNow that we know that timestamps are in minutes, maybe it would be worth to see how variations in the X axis correlate with the time of the day/minute of the hour/day of the week, to see if there's some pattern (from other visualisations it is somewhat clear that there's no correlation with the absolute time value). Maybe...", "121539": "carrDelling yes, now that we have some information from \"minutes\" we should be able to better guess what these places are." }, "source": "meta" }