Last month Muki Haklay published an interesting blog post on a digital divide in OpenStreetMap data. The basic conclusion is that more affluent areas are mapped more completely that poorer areas. Perhaps even more worryingly, the data reveal that the gap in virtual representation between affluent and deprived areas is actually growing.
These results mirror my own findings on other types of user-generated content that represents place. I've found that both Wikipedia and Google display similar patterns. Following Muki, the next step will be to use measures of affluence and deprivation (GDP, levels of internet access etc.) to more precisely examine the specific relationships.