I'm hoping that using the Microsoft Building Footprint GeoJSON files with Bing satellite will give me the best and most accurate layers upon which to now extract front yard/street/hardscape information
I don't think that's going to work, mainly because the Microsoft building footprints data is absolutely awful. See the Example: Import GeoJSON File topic, which uses the Microsoft data set. The building footprints, for the most part, aren't even close to being accurate. The Microsoft stuff is interesting from a machine learning perspective, but from a GIS perspective it's more about the terrible quality of what the algorithms produced. It's what keeps manual digitizing shops in business. If you're dead set on using the Microsoft data set, I was going to suggest using Manifold's GDAL dataport to import the Microsoft stuff, but GDAL doesn't seem to handle the Microsoft data set, so that's out. Using Manifold's native GeoJSON dataport, one thing you could do to get around the 2 GB text limit is to open the California data set in a text editor and then save it as three parts, repeating the very beginnings and very ends to correctly open and close the data portions. It's all human readable text in JSON format, so you can do that. You can then import the three parts and merge them. I think a better approach to getting building footprints is either a) use OpenStreetMap (which often has collected footprint data), or b) get data for the area of interest from whatever county it is in. For imagery, popular imagerservers like Google and Bing have a lot of issues in terms of georeferencing and orthorectification. State and local government jurisdictions often have better data based on aerial photos or drone overflights. Some locations (not sure about in California) also have NAIP photography that is sub-meter accuracy. Depending on budget, if it is just an HOA (home owner association), those tend not to be all that large, and it might be possible to get custom drone or aerial imagery shot and then mosaiced together for a reasonable price.
|