Implicit conversion from float -116.88202247191009 to int loses precision