topic Re: assess the relationship of multiple time-series variables in SAS Programming

assess the relationship of multiple time-series variables

Cruise — Tue, 07 Apr 2020 23:48:55 GMT

Hi Folks:

I'd like to assess the temporal relationship between pneumonia and minimum ambient temperature and average relative humidity over 2016 until to date. I will also have to assess the combined effected of temperature and humidity on the health outcome over time as well.

Will do the same for cold and flu as well but separately. I don't need to forecast.

My question is: What is the right procedure available using SAS to tease out these questions? proc varmax?

Thank you for your time. Screen shot of the proc print of the data and few lines of data attached just in case.

Re: assess the relationship of multiple time-series variables

ballardw — Wed, 08 Apr 2020 14:08:49 GMT

What does that pneumonia number represent? And since you say the same for Cold and Flu, what do those represent?

And since relative humidity is related to temperature you may have more complications to consider.

Re: assess the relationship of multiple time-series variables

Cruise — Wed, 08 Apr 2020 15:08:13 GMT

Thank you very much for getting back to me on the issue. I really need some help or someone to brainstorm with. The numbers for pneumonia, cold and flu represent the ratio of number of each pneumonia, cold or flu cases over the total population lived in the geographic area. RH and temp are related. Also, considering a combined effect of RH and temp is important. I was wondering about creating a simple interaction variable like: min_temp*avg_relative_humidity. I noticed that you didn't comment on the proc varmax, would you recommend proc varmax? or other proc is in consideration depending how complex the relationship under assessment, as you are hinting?