I have used 4 readily available parameters to determine this:
I have ignored pressure as this isn't something that humans notice much, and I have ignored sunshine hours, as although it is certainly useful in this context, it is not widely available.
The "typical" weather is determined from a Geometric Median of the 4 parameters. The list of stations closest to that median is then generated. I will keep a record of "scores" as per BDG's Boggy Britain thread, but I won't publish these every time, as this is meant to be a daily summary.
The median is a better description of the "typical" weather than the mean, which can be skewed by single values. Eg - if it rains only in Lerwick, then 99% of the stations will record less than the mean, but the median will be 0.0mm as long as at least 50% of stations record no rain. The Geometric Median is an extension of the median to n dimensions.
The values will have to be generated by script as the computation is far too difficult by hand, so there are unlikely to be many comments, but feel free to add them yourself.
I will start by weighting each parameter equally, although there is an argument for weighting windspeed lower than the other readings.
This post has been edited by Tim Prosser: 08 September 2011 - 23:31