Monthly data is indeed a better idea. Regarding daily data, you should check the meta data of each station about how they measured in which period and if there were changes in location, instruments or time interval. It is unlikely that you can just run any type of statistical software using that data without knowing anything about each station.