{
  "id": 5156,
  "title": "TaggedInfo from CSV files",
  "url": "/competitions/belkin-energy-disaggregation-competition/discussion/5156",
  "author_name": "",
  "post_date": "2013-07-19T17:18:31.223Z",
  "votes": null,
  "comment_count": 3,
  "views": 1633,
  "content": "<p>Hi,</p>\n<p>I'm working with the CSV data. Data from H4 wasn't giving anything good. I tried with H1 data without success. Therefore, I dug a little bit to understand what I was doing wrong.</p>\n<p>The timetick in H1_CSV\\CSV_OUT\\Tagged_Training_10_22_1350889201 runs from 1350889167.5370412 to 1350975564.598875. But, if you look at the taggedInfo, all appliances are on/off insides the window 1350943200 to 1350948900, meaning that about 7% of the entire FFT data are of interest (5700/86397). Furthermore, if you look to the data, all appliances start to a multiple of 120. I double check and it is the same for H4 data. We learn in another post that those taggedInfo was writing down by human. This is clearly impossible to have this kind of cadence for a human. Consequently, my conclusion is that the taggedInfo aren't valid.</p>\n<p>Is it possible to get valid data in CSV format?</p>",
  "messages": [
    {
      "id": "27450",
      "postDate": "07/19/2013 17:18:31",
      "content": "<p>Hi,</p>\n<p>I'm working with the CSV data. Data from H4 wasn't giving anything good. I tried with H1 data without success. Therefore, I dug a little bit to understand what I was doing wrong.</p>\n<p>The timetick in H1_CSV\\CSV_OUT\\Tagged_Training_10_22_1350889201 runs from 1350889167.5370412 to 1350975564.598875. But, if you look at the taggedInfo, all appliances are on/off insides the window 1350943200 to 1350948900, meaning that about 7% of the entire FFT data are of interest (5700/86397). Furthermore, if you look to the data, all appliances start to a multiple of 120. I double check and it is the same for H4 data. We learn in another post that those taggedInfo was writing down by human. This is clearly impossible to have this kind of cadence for a human. Consequently, my conclusion is that the taggedInfo aren't valid.</p>\n<p>Is it possible to get valid data in CSV format?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "27546",
      "postDate": "07/23/2013 13:03:13",
      "content": "<p>The CSV taggedInfo is all in fact for H1, the correct ones are in matlab</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "27572",
      "postDate": "07/23/2013 21:12:22",
      "content": "<p>Bump. Are the correct Alltagginginfos.csv loaded? I don't want to have to download the data again as it was huge just get the tagginginfos.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "27623",
      "postDate": "07/24/2013 21:19:38",
      "content": "<p>Apologies for this oversight. I am about to upload each TaggingInfo separately first, and then the entire dataset replaced with the correct info. So incase you downloaded the entire dataset, you can just download the TaggingInfo Files. I'm working to upload them within the next hour.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 27546,
      "author_name": "tonybagnall",
      "author_url": "",
      "post_date": "07/23/2013 13:03:13",
      "content": "<p>The CSV taggedInfo is all in fact for H1, the correct ones are in matlab</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 27572,
      "author_name": "domcastro",
      "author_url": "",
      "post_date": "07/23/2013 21:12:22",
      "content": "<p>Bump. Are the correct Alltagginginfos.csv loaded? I don't want to have to download the data again as it was huge just get the tagginginfos.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 27623,
      "author_name": "anirudhranganath",
      "author_url": "",
      "post_date": "07/24/2013 21:19:38",
      "content": "<p>Apologies for this oversight. I am about to upload each TaggingInfo separately first, and then the entire dataset replaced with the correct info. So incase you downloaded the entire dataset, you can just download the TaggingInfo Files. I'm working to upload them within the next hour.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "27450": "",
    "27546": "",
    "27572": "",
    "27623": ""
  },
  "source": "meta"
}