Implicit conversion from float -73.10679611650485 to int loses precision