Can we model distribution of population abundance from wildlife–vehicles collision data?