The Blackboard:

I plan to post random musings here.

Currently, I’m teaching myself, ‘R’ a statistical package. In the process, I’ll be learning more about the climate of Växjö, Sweden that anyone could possible wish to know.

I obtained a series of temperature profiles from: The European Climate Assessment Database. The data are contained in the file entitled “Daily maximum temperature TXl

I will be examining data ffor Vaexjoe, Sweden, the first station indicated on the list. The data set contained daily temperatures collected between 1918 and 2001. Meta-data include two types of error flags indicating either total instrument failure or suspicious data.

I imported the data into ‘R’. After eliminating the data corresponding to pure instrument failure, I created this histogram.

Temperatures are indicated in tenths of a degree C, so an indication of 100 means 10C.

The histogram suggests temperatures in Vaexjoe is bi-normally distributed. One peak is near 0C, the freezing temperature of water. The other peak is at a somewhat warmer level.

What can I say about any statistical analysis at this point?

Distinctly bi-modal data present a confounding factor. Statistical analyses will require some care, particularly with regard to the assumption that data are normally distributed.

Next Post:
The R manual & What’s in it. »