When it comes to “global temperature data sets” I have always reached for some strong adult beverages before, during, and after reading any conclusions drawn from them.
Why?
First, the data collection points, i.e. weather stations, are NOT evenly distributed across the globe. Don’t think there is much on the ground reporting from sites along the Trans Siberian railroad or Central Gobi Desert do you?
Second, I doubt that any of the temperature recording devices have been initially calibrated to the nearest 1/10th degree C. Nor, have they been maintained with regular recalibrations.
Third, There was no need to report and record temperatures with any degree of accuracy until the age of flight, the mid 1930s. Only then did accurate temperature readings become, potentially, a mater of life or death.
Fourth, I am still waiting to see ice bergs in the Gulf of Mexico as breathlessly reported in the early 1970s. This was supposed to happen because all temperature records then available pointed to a severe cooling trend.
Interestingly enough, there were many more data collection points during the Cold War. The Soviets did lots of measurements in Siberia and elsewhere. Other countries did too. But as the old Soviet Union collapsed the weather recording was abandoned for the most part, except, in particular, in big cities. Thus the Siberian temperature trend started going way up. Not just because Siberia got warmer, it did, verified by satellite. But also because the cities warmed faster than the prior small town and rural stations. So part of "global warming" is an artificial warming trend due to loss of temperature recording at smaller population centers.