The three-line MATLAB script that plotted this graph is clim65.m. You also need all the data, which is contained in the script load_Period_HadCRUT4_CO2_TSI.m, along with the moving-average function ma.m. (In some other systems, e.g. R, ma is part of the library, but strangely not in MATLAB.)
Just those three files is all you should need: put them in the same folder and run clim65 from the command window. If you encounter any difficulties feel free to email me at my surname (all lower case) at cs.stanford.edu.
Caveat: TSI is based on NOAA's reconstruction. Stanford's Leif Svalgaard has challenged its accuracy and has proposed an alternative reconstruction. If you use it instead of NOAA's TSI data you must increase CO2 climate sensitivity from 1.7 to 1.87 degrees per doubling of CO2, and solar sensitivity from 1.0 to 1.83 degrees per W/m2 of top-of-atmosphere TSI in order to demonstrate the removal of all climate impacts save CO2 and Sun with a 65-year filter. This might motivate non-luke-warmers to take a look at Svalgaard's reasons for preferring his reconstruction.
The Excel spreadsheet is here. It assumes recent versions Excel for safety since older versions were more vulnerable to viruses in spreadsheets you might run across on the web. If there's interest I could save a version that works with Excel'97 but you'd want to be sure it was from me and not a copy someone's been playing with.