Here is the dataset containing CSVs of all the train+test paths from all the 24 sites we are asked to predict from.
\nIn each CSV, each row depicts a particular timestamp(index), with all the 52 columns of information from all 10 datatypes at that timestamp. As is known, accelerometer, gyroscope, magnetometer, rotation vector, and their uncalibrated values are available in abundant, but wifi/iBeacon values are more rare, so if any value is not known at a particular timestamp, then the cell has been left empty, you may wish to fill those with the mean or check if other path files have a wifi/iBeacon reading for that position, or maybe the last seen values, whichever you see fit.
\nTrain files are named like this:
\n\n\nsiteID_floor_path.csv
\n
Test files are named such:
\n\n\nsiteID_path.
\n
This notebook was used, if you wish to edit some parts or get the train files for all sites, although keep in mind some of them(about 15 sites~) are formatted differently/poorly so you will probably run into errors.
\nHope it helps, have fun!
", "messages": [ { "id": 1184706, "postDate": "2021-02-03T17:31:49.417Z", "content": "Here is the dataset containing CSVs of all the train+test paths from all the 24 sites we are asked to predict from.
\nIn each CSV, each row depicts a particular timestamp(index), with all the 52 columns of information from all 10 datatypes at that timestamp. As is known, accelerometer, gyroscope, magnetometer, rotation vector, and their uncalibrated values are available in abundant, but wifi/iBeacon values are more rare, so if any value is not known at a particular timestamp, then the cell has been left empty, you may wish to fill those with the mean or check if other path files have a wifi/iBeacon reading for that position, or maybe the last seen values, whichever you see fit.
\nTrain files are named like this:
\n\n\nsiteID_floor_path.csv
\n
Test files are named such:
\n\n\nsiteID_path.
\n
This notebook was used, if you wish to edit some parts or get the train files for all sites, although keep in mind some of them(about 15 sites~) are formatted differently/poorly so you will probably run into errors.
\nHope it helps, have fun!
", "rawMarkdown": "[Here](https://www.kaggle.com/aatiffraz/indoor-location-navigation-tabular-data) is the dataset containing CSVs of all the train+test paths from all the 24 sites we are asked to predict from. \n\nIn each CSV, each row depicts a particular timestamp(index), with all the 52 columns of information from all 10 datatypes at that timestamp. As is known, accelerometer, gyroscope, magnetometer, rotation vector, and their uncalibrated values are available in abundant, but wifi/iBeacon values are more rare, so if any value is not known at a particular timestamp, then the cell has been left empty, you may wish to fill those with the mean or check if other path files have a wifi/iBeacon reading for that position, or maybe the last seen values, whichever you see fit.\n\nTrain files are named like this:\n\n> siteID_floor_path.csv\n\n Test files are named such: \n> siteID_path.\n\n\n[This](https://www.kaggle.com/aatiffraz/converting-files-to-csv) notebook was used, if you wish to edit some parts or get the train files for all sites, although keep in mind some of them(about 15 sites~) are formatted differently/poorly so you will probably run into errors.\n\nHope it helps, have fun!", "votes": 79 }, { "id": 1238208, "postDate": "2021-03-14T18:32:42.007Z", "content": "Hi @aatiffraz, thank you for your dedicated work. However, while I am studying your work, I found that for the Wifi features (BSSID, SSID, etc.), at one timestamp, there are many Wifi signals, but in your CSV type data, only the last signal is recorded. I am not quite sure this is your intention or it is a bug? I would appreciate it if you could confirm it. If this is a bug, maybe another warning thread might be useful for those who are using this dataset.
", "rawMarkdown": "Hi @aatiffraz, thank you for your dedicated work. However, while I am studying your work, I found that for the Wifi features (BSSID, SSID, etc.), at one timestamp, there are many Wifi signals, but in your CSV type data, only the last signal is recorded. I am not quite sure this is your intention or it is a bug? I would appreciate it if you could confirm it. If this is a bug, maybe another warning thread might be useful for those who are using this dataset.", "votes": 3 }, { "id": 1210667, "postDate": "2021-02-19T15:39:54.270Z", "content": "It can be really useful! Thank you for your hardwork!
", "rawMarkdown": "It can be really useful! Thank you for your hardwork!", "votes": 1 }, { "id": 1221871, "postDate": "2021-03-01T11:09:17.910Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1191968, "postDate": "2021-02-08T20:22:15.213Z", "content": "It seems very helpful, thanks! 👍
", "rawMarkdown": "It seems very helpful, thanks! 👍", "votes": 1 } ], "comments": [ { "id": 1238208, "author_name": "Minh Tri Phan", "author_url": "", "post_date": "2021-03-14T18:32:42.007000", "content": "Hi @aatiffraz, thank you for your dedicated work. However, while I am studying your work, I found that for the Wifi features (BSSID, SSID, etc.), at one timestamp, there are many Wifi signals, but in your CSV type data, only the last signal is recorded. I am not quite sure this is your intention or it is a bug? I would appreciate it if you could confirm it. If this is a bug, maybe another warning thread might be useful for those who are using this dataset.
", "votes": 3, "replies": [] }, { "id": 1210667, "author_name": "Dongkyu Kim", "author_url": "", "post_date": "2021-02-19T15:39:54.270000", "content": "It can be really useful! Thank you for your hardwork!
", "votes": 1, "replies": [] }, { "id": 1221871, "author_name": "", "author_url": "", "post_date": "2021-03-01T11:09:17.910000", "content": "", "votes": 0, "replies": [] }, { "id": 1191968, "author_name": "Sinan Calisir", "author_url": "", "post_date": "2021-02-08T20:22:15.213000", "content": "It seems very helpful, thanks! 👍
", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "1184706": "[Here](https://www.kaggle.com/aatiffraz/indoor-location-navigation-tabular-data) is the dataset containing CSVs of all the train+test paths from all the 24 sites we are asked to predict from. \n\nIn each CSV, each row depicts a particular timestamp(index), with all the 52 columns of information from all 10 datatypes at that timestamp. As is known, accelerometer, gyroscope, magnetometer, rotation vector, and their uncalibrated values are available in abundant, but wifi/iBeacon values are more rare, so if any value is not known at a particular timestamp, then the cell has been left empty, you may wish to fill those with the mean or check if other path files have a wifi/iBeacon reading for that position, or maybe the last seen values, whichever you see fit.\n\nTrain files are named like this:\n\n> siteID_floor_path.csv\n\n Test files are named such: \n> siteID_path.\n\n\n[This](https://www.kaggle.com/aatiffraz/converting-files-to-csv) notebook was used, if you wish to edit some parts or get the train files for all sites, although keep in mind some of them(about 15 sites~) are formatted differently/poorly so you will probably run into errors.\n\nHope it helps, have fun!", "1238208": "Hi @aatiffraz, thank you for your dedicated work. However, while I am studying your work, I found that for the Wifi features (BSSID, SSID, etc.), at one timestamp, there are many Wifi signals, but in your CSV type data, only the last signal is recorded. I am not quite sure this is your intention or it is a bug? I would appreciate it if you could confirm it. If this is a bug, maybe another warning thread might be useful for those who are using this dataset.", "1210667": "It can be really useful! Thank you for your hardwork!", "1221871": "", "1191968": "It seems very helpful, thanks! 👍" } }