Facebook Academic Research API: Early Access Now Available

Facebook Launches New Research API for Academic Teams

A select number of academic research groups are now receiving access to a novel tool developed by Facebook, designed for the aggregation of extensive, real-time data from the world’s largest social media platform.

Addressing Past Concerns

Meta, formerly known as Facebook, continues to navigate the repercussions of the 2018 Cambridge Analytica incident. This event involved the unauthorized harvesting of personal data from millions of Facebook users by a political consulting firm, used to construct detailed voter profiles.

Following this scandal, the company deactivated thousands of APIs over the subsequent three years. Now, Meta is initiating a phased restoration of broader data access specifically for academic research purposes.

Introducing the Researcher API

TechCrunch received a preview of Facebook’s new academic research API and interviewed Kiran Jagadeesh, Facebook Product Manager and leader of the project within the Facebook Open Research & Transparency (FORT) team.

Jagadeesh characterized the Researcher API as a beta iteration of a more comprehensive toolkit the company intends to offer, stating, “This represents merely the initial phase.” The API, initially unveiled at F8 earlier this year, is built on Python and operates within JupyterLab, an open-source notebook environment.

Controlled Access and Data Security

Recognizing past privacy issues, the new Researcher API is being rolled out with specific safeguards. Initially, access is limited to a small cohort of established academic researchers through an invitation-only process.

Meta plans to broaden access in February 2022, integrating feedback from the initial trial phase into a wider academic launch.

To further protect user privacy, the API operates within a secure, controlled environment described by Jagadeesh as a “digital clean room.” Researchers can access the environment via a Facebook VPN, analyze data, and perform calculations, but the raw data itself cannot be exported – only the resulting analyses.

This approach aims to prevent re-identification of users, although some critics may argue that the data the API collects is already publicly available, albeit difficult to aggregate and analyze using existing Facebook tools.

Data Availability and Scope

At its launch, the API will grant access to four primary categories of real-time Facebook data: pages, groups, events, and posts. Data will initially be sourced only from public sources within the United States and the European Union.

For groups, events, and pages, at least one administrator must be located in a supported country for the data to be accessible through the API.

Analytical Capabilities

The tool enables researchers to analyze substantial volumes of text data using techniques such as sentiment analysis, which assesses the emotional tone and valence expressed in online discussions.

Alongside text-based posts, researchers can also access supplementary information like group and page descriptions, creation dates, and post reactions.

However, multimedia data like images, as well as user demographic information (age, gender, etc.), will not be included. Data from Instagram will also not be available initially, though the team is exploring potential integration in the future.

Collaboration and Future Development

The FORT team intends to collaborate closely with academic researchers to refine and expand the current tools, acknowledging that the API is still under development.

Meta has extended invitations to researchers from 23 academic institutions globally to participate in initial testing.

Access was granted on November 15th to researchers who have completed the onboarding process and agreed to the stipulated privacy policies, which include a prohibition on re-identifying individuals within the data.

Expanding Access and Transparency

While currently limited to academic institutions, the FORT team plans to explore extending access to other groups, including journalists.

The team aims to establish a public roadmap, providing researchers and journalists with a transparent view of ongoing development efforts.

Rebuilding Trust

Meta faces the challenge of rebuilding trust within the research community. In August, the company revoked access to advertising data for researchers from NYU’s Cybersecurity for Democracy project, which focused on tracking misinformation and political ads. This action drew criticism from academics and regulators.

In September, Facebook apologized to the Social Science One research group for providing incomplete data, which compromised months of work and analysis.