Did you first formulate a hypothesis, then test it against the data? Or, as the description of your methods suggests, did you trawl through a bunch of weather data and a bunch of ephemerides, and see what popped out?