Hi Alicia, Just saw your question.
Like most applications, the resolution and scale of the imagery depends on what features you are interested in. Landsat and Sentinel (30m) imagery is great if you want the amount of hard surfaces in an urban area like a city. This will give data at the neighborhood scale, and is suitable for landcover projects.
If you want more detailed features at the landuse scale, such as individual houses, sidewalks and driveways, industrial and commercial complexes, etc., then you need higher resolution satellite data such as Worldview-3 (50cm or 2m) or Hexagon 30m. Of course you want the imagery to be multispectral, and the infrared band is required to separate vegetation from man-made features.
A rule of thumb is that the smallest feature you can "reliably" detect is 3 x 3 pixels. So if you were using 50cm WV3 data, that would be a feature 1.5m x 1.5m, or 2.25 square meters.