Within the location theme of the Mobile 2.0 project, we’ve been producing quite some mash-ups on top of existing services like foursquare. Using existing services to build apps certainly has advantages in terms of development burden, and can leverage their popularity and existing user bases. Using services as materials however does have consequences that go beyond the technical realm, and are not necessarily about the API alone – especially in a research context.
Using mashups is research means being dependent on other services, and publicly releasing mashups means giving up the control you may have in smaller-scale studies. APIs change, services go down, and you might have to adapt the experience you were planning to offer. A service is more than just its API; they are part of businesses, with goals, PR strategies – and people. Services have target audiences, and have usage cultures that should fit your purposes.
In short, using existing services as materials has been very useful for us, but does require understanding your material. More about the various considerations in using mash-ups for research and our experiences in this position paper.