Implicit conversion from float -114.46685878962535 to int loses precision