Implicit conversion from float -118.17073170731706 to int loses precision