In regards to your first question, one commonly applied technique for measuring the extent of human impact, whether it's of agriculture, urbanization, or other human land use change, is to apply the human footprint map developed by Sanderson et al. 2002. This spatial layer effectively measures the relative human influence in every biome of earth's surface. If the goal is to understand how urban areas differ from natural ones in terms of species richness, then values from the human footprint layer can be used as an explanatory variable to predict richness in these two respective areas. Another more recent attempt to classify human footprint can be found by Venter et al. 2016: https://www.nature.com/articles/sdata201667.
I had planned to do the inventory of species in smaller areas like Gardens and parks, However my area has some natural areas like large Hilly areas and river with plenty of vegetation; which i feel would be not feasible to do counting and do inventory, Hence planned of doing random Sampling. what are your suggestion about this?