Implicit conversion from float -120.22443890274315 to int loses precision