Implicit conversion from float -118.13148788927336 to int loses precision