Implicit conversion from float -113.04182509505702 to int loses precision