Statistics Charts Questions



  • Avatar
    James Downing

    I have also noticed what may be a bug, the charts are so useful, but they have some weird viewing increments.  It makes it seem like it will display data in any increment of time, but it will only display the "detail" graph view window up to 5 hours in length.  Then it won't display any time of window between 5 and 13 hours.  After 13 hours, it moves to the "smoothed" graph view.  Annoying.

  • Avatar
    Steve Rossi

    I'm just another user like you, but I can answer some of your questions. 

    1) This is ambient light. I don't know the units or range of measurement. I asked a few months ago in this forum since I wanted to use some custom coding to determine room occupancy based on known ambient light from outside sources and when (artificial) lights are on in a room.

    2) The puck takes a sensor reading every 25 seconds. By default it only views the last 150 readings (per API settings.. this used to be 300 in July) which is about 62 minutes. I've found that this does match what I see on the homepage, but not necessarily what I see on the display of the puck. Once you change the reporting time from default I have no idea how the system aggregates which values to show. There is a column in the API data for offset, which I see populated on some rows. I don't know how this factors in.  I'd encourage you to get api access to download the data locally to make better comparisons.  I'll warn you though, the data collection is significant -- I have 3.2 million rows for 8 pucks over the last 4 months.

    3) I've seen this with some of my vent sensors. The airflow initially out of the ducts is warmer due the heat of my attic. This was more pronounced the longer the period went between the fan cycling.  something to check for is if you have any airflow. I've seen some houses where due to pressure differentials between rooms/floors there is slight positive pressure coming from the registers. You could just be seeing the result of very slow moving air warming up from the attic.

    4) I have no idea what this indicates. Out of 2.4 million data points 800k (36%) for me show 3000. Min=718, Max = 65535, Mean = 1826.I can't see any patterns in my data.

    5) The vents are capable of 50% when manually setting them. Automatic mode at this time only utilizes open or closed. Expect to see some randomness with 1% and 99% values. 


    I've thought about trying to feed all the data into a RRDtool database to create graphs based on time periods instead of number of records to get a better idea. Overall the amount of data is impressive, but I have yet to figure out a better way of interpreting it in a meaningful way. Personally I want to see how good of a job the system did at maintaining a temperature in a room or where it ended up dumping cold/hot air from the system as part of the decisions. It's definitely helped with temp variations in my two story home, but it also closes vents at random times that I can't figure out why.



  • Avatar
    James Downing

    Thanks for the reply!  It sounds like I'd enjoy getting API access.  I'll have to look into it.  I only have Flair in one room, a windowless basement, and was hoping it would help prevent overcooling of the space in the summer or overheating in the winter.

    Follow ups:

    2. I felt like there was a smoothing function that gets applied to the reported number, but at present that number matches the graph exactly, so I can't prove it.

    3. The ducting is between two conditioned spaces, not in the attic, but that could have made sense.  It doesn't make sense for it to stabilize at 80 when the rest of the house is about 68 degrees.  See below.

    5.  This is a disappointment, it could utilize a pretty simple PID that could infinitely vary the open % (might also make them less noisy).

Please sign in to leave a comment.