Data available in both text and csv formats at the NASA, GISS website have been routinely cited as indicative of global warming despite their known weaknesses.  The three years 2015 through 2017 are widely reported as the three hottest years on record.

Tony Heller, however, has demonstrated that tampering with data from the U.S. Climatology Network (USHCN) has created the illusion of much higher temperatures in reported data than in the original data, for the continental United States.  This leads one to wonder how much not so widely known "adjustments" in GISS data have been responsible for similar results at a global level.

The GISS data are updated around the middle of each month, and I have compared the January and March versions in figure 1, for the years 1881 through 2017.  The data are smoothed over two years, in that, for example, the 1881 data point is the average of 1880 and 1881 and 2017 the average for 2016 and 2017.  This is commonly done to make data more presentable, allowing movements to be more clearly discerned and to smooth out the effects of "abnormal" years.

This comparison allows two interesting observations.  Firstly, the March revised data indicate somewhat higher temperatures in roughly the early third of the period, and somewhat lower temperatures the last third of the period.  The overall pattern, one of warming, however, remains...

...albeit slightly more muted than in the January version of the data, with 2015 through 2017 still being indicated as record hot years.  Importantly, the data points for 2016 and 2017 are much lower in the revised data, and the three years 2015 to 2017 show no acceleration in temperatures.

The second observation is that the 2017 data point (an average of 2016 and 2017 anomalies) is slightly lower than the 2016 point (an average of 2015 and 2016).  This implies that temperatures have fallen substantially during some months in 2017 below those of 2016, as if some sort of cycle may be taking place.  One can observe this from the monthly data in Figure 2, where temperatures peak in the first three months of 2016 before declining sharply in April and falling below those of 2015 from October through December.  Temperatures for the first eleven months of 2017 are below those of the corresponding months in 2016.  Importantly, the downward trend in temperatures continues into 2018, where January and February anomalies are both below December 2017, and those of the previous three years.

At least part of the cycle in the record temperature anomalies begun in late 2015 is related the "El Niño" effect and was recognized as making 2016 the "hottest" year on record, as early as July of that year, by Gavin Schmidt, director of NASA's Goddard Institute for Space Studies (GISS).  El Niños typically occur every three to five years, with related temperature effects often lasting for many months.

Of course, any data set from only 1880, and inadequately covering the earth's surface area, does not provide a definitive answer to the question of "global warming" in terms of geological time of thousands of years, and representing the entirety of the Earth.  But, as it has been the data set often referenced to substantiate global warming, one would have thought the substantially lower temperatures of the last many months would have merited highlighting in the mainstream media.