Sean Foo (designer of the FTS API) points out a couple complexities of their system that are worth noting:
Years can be associated with various entities in their system. For example, contributions can have a given year, but pooled funds can also have a given year. It's possible that CERF-2012 funds end up in the 2013 amounts for a given emergency. Or a non-pooled-fund contribution could be made at the end of 2012, for use in 2013.
Contributions to an emergency don't necessarily have to match the country of that emergency. For example, a contribution to the Syrian emergency could be for refugees now in Jordan.
These likely aren't major (in spot checking, the numbers come out very close between the indicators produced by the existing code and those from FTS reports), but should probably be resolved at some point. Presumably the FTS reports have the desirable behavior?
Looking things over, I'm not sure the current API can give the same numbers as the FTS reports. Perhaps it could be changed, or perhaps we can live with these small discrepancies?
Sean Foo (designer of the FTS API) points out a couple complexities of their system that are worth noting:
These likely aren't major (in spot checking, the numbers come out very close between the indicators produced by the existing code and those from FTS reports), but should probably be resolved at some point. Presumably the FTS reports have the desirable behavior?