Implicit conversion from float -47.418086500655306 to int loses precision