I am writing a masters thesis about estimating solar potential based on open source resources. Part of my methodology is to use a sample of buildings to estimate solar potential for the whole city. For that purpose, I need to know how representative the buildings I extracted from OSM are for all the buildings in city.
Therefore, I was wondering how can I trace back why these buildings were selected to be digitized on OSM?
OSM is a collaborative geographic database, fed by the goodwill of volunteer contributors. Therefore there are no real input specifications (as there might be in a reference database), just guidlines, which you can find on the project wiki. It is therefore difficult to answer the question: "why was this building designed like this?", because it is something that will depend on the contributor, their way of working, their interests, etc.
However, a large proportion of the buildings that are now in OSM come from automated imports. This is for example the case in France, where most of the buildings come from an automated import from the cadastral survey. In this case it is easier to identify the reason for the entries, since it corresponds to the specifications of the imported data. So if you are able to distinguish between buildings that have been entered manually and those that have been imported automatically this will already be a good step in qualifying the representativeness of the (imported) buildings.
Of course, this will not answer all the notions that can be included in the term "representativeness". For example, it is difficult to know how up-to-date the data are, unless you compare them with aerial or satellite images. But it should already answer some questions.
