Partnerships

To learn about how to access Facebook data, see here; for a Dataverse of our Facebook datasets, see here.

Social Science One develops models for industry collaboration, leads demonstration projects with major companies, and helps others develop their own partnerships. Only a few of these partnerships are run by Social Science One or housed with us; most of the work we do involves sharing information so academics worldwide can fan out and make partnerships with different parts of the corporate world. Throughout this hybrid process, we have facilitated sharing vast amounts of data including some of the largest social science datasets ever constructed. We have also seen many corporate employees detailed to joint projects and large amounts of expertise flowing between academia and industry.

Currently, Social Science One is helping to structure partnerships with Facebook and Microsoft, representing several different models of collaboration. With Facebook, we put together an RFP-based data sharing system, overcoming considerable legal, institutional, and political barriers to release a dataset of all URLs shared on Facebook (now totalling more than 17 trillion cell values). We also help make data available through the Facebook Ad Library and CrowdTangle APIs, and have initiated and support a partnership to study the 2020 election.

We have a collaboration with Microsoft to build SmartNoise software and to aid our OpenDP project which seeks to engage a community of collaborators in academia, industry, and government to build trustworthy, open-source software tools for privacy-protective statistical analysis of sensitive personal data. These tools offer the rigorous protections of differential privacy for the individuals who may be represented in confidential data and statistically valid methods of analysis for researchers who study the data.

In the recent past, our team worked in partnership with Crimson Hexagon (now Brandwatch) to download and analyze all Chinese social media posts before the Chinese government could censor them. The team has also partnered with 48 media companies to perform a massive experiment that enabled researchers for the first time to randomly assign what news outlets would publish and when in order to measure the effect of news media exposure on political behaviors.