This program analyzes temperature data from HadCRUT data set.
It is easy to say that temperature is rising or dropping if you can select the years what you compare. This program shows all the meaningful year ranges that are available in the measurements.
Meaningful data ranges are considered all those, who are 7 - 50 years. The whole data set is 168 years.
In the data set there are currently years 1850 - 2018. At those years there are about 1400 year ranges between seven and fifty years that show statistically meaningful increase in temperatures. Additionally there are about 1000 year ranges that show increase in the temperature but it is not statistically important. There are also about 1000 year ranges that show non-statistically-important cooling. But there are still about 200 year ranges that show statistically important cooling.
When we put the ranges to the chart, we get a better view of oscillation of warming and cooling periods. The chart hides outliers nicely.
Range lines are randomly spread between the range-length-year and the next range-length-year to avoid them totally shadowing each other. Statistically most significant range is on the top.
In the bottom we see that there is oscillation of little over 20 years between cooling and warming periods. In the middle there seems to be larger oscillation of about 60 years but we really see only two cooling periods that does not make it possible to make any conclusions. Of the periods over 40 years we do not have enough measurements to say anything meaningful. If there is oscillation of 60 years, we would not see it in the top of the chart yet.
Program loads data set directly from UK Met Office. It is implemented with R statistical language and it does not need any extra libraries.