Implicit conversion from float -116.94915254237287 to int loses precision