Implicit conversion from float -122.72040302267004 to int loses precision