Implicit conversion from float -122.31979030144166 to int loses precision