Here I find yearly means from monthly data, using xarray, then I use linear regression (machine learning) to plot a line of best fit, then the model predicts the temperature in 2100. Please note that all temperatures here are in Kelvin.