The concept of habitat is a great example of the messiness and complexity of nature. In general, a habitat is an area where a species or a group of species live. What makes a habitat is a mix of numerous factors such as geology, weather, interaction with other species, soil chemistry, shade etc.
A defined habitat is essentially a model which can help predict what you can observe in an area. Some models are better than others. The one I find least useful is the Field Guide to Forest Ecosystems of Central Ontario. This one is limited to forest ecosystems. The problem is the huge overlap between each ecosystem and vegetation type described. It does do a good job of predicting the mix of common vegetation found at each ecosystem and vegetation type described.
Another thing that can be disappointing are books that describe habitat types. Because of the geographical scope of these books, the number of habitats described are limited.
More interesting is the plant community list from the Natural History Information Centre in Ontario. It lists 516 different plant communities. While lots of the plant communities overlap a lot, it lists many uncommon or rare plant community types. If you can find these uncommon or rare plant community types you are bound to find some uncommon or rare species.
A neat perspective on habitats can be found in the Ecomaps chapter of the recently published Field Guide to the Woody Plants of the Northern Forest by Jerry Jenkins. On page 54 he shows ecological patterns at four scales – regional (100 km), physiographic (10 km), landscape (1 km) and species (2 m). I personally think it is useful to extend this in both directions. On a larger scale, you run into the concept of ecozones. In Ontario these zones are Hudson Bay Lowlands, Ontario Shield and Mixedwood Plains. These ecozones are part of a hierarchical system – ecozone, ecoregion, ecodistrict, ecosystem, ecosite and ecoelement. These are described at https://www.ontario.ca/page/ecosystems-ontario-part-1-ecozones-and-ecoregions#section-2. You can extend ecological patterns in the opposite direction. A species can be a habitat for another, as a host. The host can be exploited by parasites, parasitoids (parasites that kill hosts) or simply as a substrate (e.g. lichens). Hyperparasites can prey on parasites and hyperparasitoids can prey on parasitoids. Other combinations are possible too. There might even be parasites of hyperparasites that can be seen with the naked eye. Once you go microscopic the levels can go even further.
The taxonomic group you study can effect how you see the landscape. Someone into dragonflies would be looking for different types of waterbodies. Riffles are quite important for finding uncommon species like the snaketails. Butterflies like open sunny areas generally, but also occur in shady forests and bogs. Butterflies are hard to search for by habitat though. With birds, an important habitat is the edges of large lakes where birds land exhausted during migration, especially during inclement weather with winds in the correct direction. With lichens, you look for areas with less pollution, less trampling and older trees. With insects there are tons of microhabitats (under rocks, tree trunks, grass, etc) and trapping methods for finding them (sweep netting, beating, light traps, baits, barking, etc.) .