Implicit conversion from float -71.160409556314 to int loses precision