coornet.org

What It Is

CooRnet is an R package designed to detect coordinated link sharing behavior (CLSB). It takes a set of URLs and their posting times, identifies whether public entities such as Facebook pages, groups, and verified public profiles repeatedly share the same news links within unusually short time windows, and outputs a network of the entities involved. The approach is mainly intended for research into social media manipulation, problematic information diffusion, election communication, and public-issue discourse.

Core Features and Ecosystem

Its core algorithm works in two steps: first, it estimates a time threshold to determine whether multiple distinct entities shared the same URL within a time span that is unusually short relative to the overall dataset; then it groups entities that repeatedly share the same news items within that coordination window. Outputs include a data frame with an is_coordinated flag, an igraph network of coordinated entities, and exportable CSV and GraphML files. Higher-level functions can also generate component summaries, identify top URLs in coordinated networks, surface top shared posts, and sort results by metrics such as engagement.

Language, API, and Integration

CooRnet is clearly aimed at R users and is installed from GitHub via devtools. It relies heavily on a CrowdTangle API key to retrieve public Facebook/Instagram sharing data. The initial URLs can come from MediaCloud, Twitter APIs, GDELT, or CSV files. Optional integrations include the OpenAI API, using GPT-3.5-Turbo to generate English descriptive labels for clusters, and the NewsGuard API, which returns average ratings for domains shared by coordinated networks. Overall, it is more of an analytical component in a research toolchain than a full SaaS platform.

Pricing and Deployment

The source text does not mention any fee for CooRnet itself. It is installed as a GitHub R package and is suitable for running in a local R environment. Note that a CrowdTangle key is required, while OpenAI and NewsGuard are optional external services; NewsGuard requires an active subscription. The text does not specify the license, CRAN release status, or commercial support.

Pros, Cons, and Best-Fit Users

Its strengths are a clearly defined research problem, well-explained algorithm, examples covering the workflow from data collection to export, and references to related papers, making it easier to reproduce academic work. Its drawbacks are its heavy dependence on external APIs—especially the availability of CrowdTangle—and the fact that it is only suitable for users familiar with R and social media data. It is a good fit for research teams in communication studies, computational social science, fact-checking, and platform governance, but not for general operations teams looking for an out-of-the-box dashboard.

Access from China

The source text does not provide information on access, payment, or mirrors for users in mainland China. In practice, using it may involve external services such as GitHub, OpenAI, CrowdTangle, MediaCloud, and NewsGuard, so network connectivity and account permissions may be uncertain. Possible alternatives include building custom analysis workflows with igraph/NetworkX, using Gephi for visualization, or working with GDELT or MediaCloud Explorer.