The date should be easy enough to get: grep the meta-tags array for the first item where type=>"DATE" and then take the "name" value from it. But on the other, I don't know how your program is supposed to know which type=>"COMPETITION" entry to use, so it could come up with either "Qualifiers" or "World Cup 2014" as the value. Unless there's a pattern to the data that you can count on -- the sport always comes last of any COMPETITION entries, for instance, or it's always the one with the smallest "id" value -- you're going to have a hard time picking it out programatically.