Abstract: Developments at the intersection of the social sciences, computer science, and the Internet have opened up new vistas for studying social systems. These opportunities often come with substantial start up costs. For example, the Internet enables experiments at larger scale/lower costs than was previously conceivable. However, the start up cost for managing/coding online experiments can still be substantial. I will discuss two data infrastructures that my lab has been working on. The first is Volunteer Science (www.VolunteerScience.com). Volunteer Science is a platform which enables easy management of online experiments, integrates with existing infrastructures, such as Qualitrix and Mechanical Turk, and reduces coding overhead for experiments that require synchronous communication amongst groups. It also offers a large, international, volunteer sample of subjects. The second infrastructure is an open, disambiguated version of political contribution (FEC) data. Political contribution data are incredibly rich and complex, spanning 36 years, with geographic, occupational, temporal, and employment information. However, their scale and their messiness—most notably, absence of unique identifiers for individuals—have impeded scientific progress. We will be offering an open methodology (as well as a published data set with unique identifiers—with early access to seminar attendees) for disambiguating political contribution data, which I will outline in this talk.