Thanks for the good question!
Indeed, there are differences, mainly due to two factors:
1) EPH uses only experimental information, and most solids do not have a measured thermodynamic data. The Materials Project uses a combination of experimentally thermodynamic data for aqueous ions with computed enthalpies of formation for the solid states. Hence, the Materials Project has an order of magnitude more data on solids than any other resource. Please see the paper on Pourbaix diagram formalism for more details.
2) That being said, one area where the Materials Project may have less accurate data is for hydroxides (which is one reason for the differences between the EPH and the Materials Project Cu diagram); we tend not to have good information on the proton positions, which can cause errors in the energy calculation of these compounds.
I hope this helps!