Implicit conversion from float -112.46478873239437 to int loses precision