Dears all,
I am looking for methodologies allowing to generate hourly temperature data from a 3-hours dataset. Do you know about suitable methodologies on this regard?
Many thanks for your support.
Best,
Giorgio
You could try interpolation. But my advice would be: don't do it. If you only have data for every three hours, that data is not precise enough to know what happens every hour.
As I see it you have two options
1) aggregate the other variable(s) of interest to the 3 hour time window so that all the variables are comparable
2) get from somewhere an auxiliary variable that give the needed hourly resolution and use that to interpolate. For example somewhere that does have the hourly data that you can aggregate to the 3 hours interval and build a model of it. The problem is that the hourly will be more variable while the 3 hourly will be more homogeneous - smoothed out; and I am presuming that is these extra peaks and troughs that are of interest Interpolating that meaningfully will be difficult as the relationship may well be different at different times of day, different seasons etc.
- Badges
- Science topic