2
$\begingroup$

I have a dataset that gives values of

  1. temperature,
  2. Relative humidity
  3. Geopotential
  4. Ozone mass mixing ratio
  5. Potential vorticity
  6. Fraction of cloud cover
  7. Specific cloud ice water content
  8. Specific cloud
  9. liquid water content
  10. Specific humidity
  11. Specific rainwater content
  12. Specific snow water content
  13. Divergence
  14. U-component of wind
  15. V-component of wind.
  16. Vertical velocity

which parameters will make sense to do regression for electricity consumption? In search of this, I got this paper, https://www.sciencedirect.com/science/article/pii/S0306261914005674 which use temperature and relative humidity for the same, and there are other papers also which uses temperature only as a parameter. Do you think that any other parameter from above I should incorporate? Can anyone please give an answer to the above question and pointwise reasoning for each item parameter and why it is used to calculate electricity consumption?

Improve this question
$\endgroup$
3
  • 1
    $\begingroup$ I would think the main influence in power consumption is temperature (for air conditioners/heaters/refrigeration). Cloudcover fraction could affect light usage. RH/U and V winds/specific humidity will be factors in how the temperatures feels so also likely factors. In theory people may sleep a bit more on rainy days, but also may do more things indoors, so that's an interesting one, so I'd think 11/12 are involved. There are other indirect connections (3/5/13/16 are factors in cloud coverage and rain), clouds affect temperature, etc. $\endgroup$
    – JeopardyTempest
    Commented Apr 25, 2022 at 13:31
  • 1
    $\begingroup$ Not sure what altitude ozone you're talking. High ozone at low levels is likely indicative of sinking air/high pressure... so another indirect one. But I think basically it all comes down to 1) climate control usage 2) lighting and 3) whether people are home/more active (though a lot of electricity usage can be industrial too, which could affect things). In the end, if it's easy, just toss them all in, why not. If you want the most likely candidates, I'd think 1, a combined magnitude of 14/15, 10, 6, and maybe 11/12 (if they show precipitation, I'm not familiar with the terms) in that order. $\endgroup$
    – JeopardyTempest
    Commented Apr 25, 2022 at 13:35
  • $\begingroup$ I agree with @JeopardyTempest. If you have the data & the time compare temperature with apparent temperature, there may or may not be a difference. Apparent temperature considers temperature, humidity & wind speed. $\endgroup$
    – Fred
    Commented Apr 25, 2022 at 19:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.