Skip to content

CooRTweet (with CooRnet predecessor)

FreeInstitutional

Publisher: github.com/nicolarighetti/CooRTweet | Type: Non-detector | Cost: Free

TL;DR

A free, open-source R package by Righetti and Giglietto that detects coordinated link-sharing and behavioural networks across social-media datasets, with native support for Meta Content Library data. The R-native alternative in the institutional CIB ladder for analysts already inside the IFCN-signatory MCL gate.

What it does

CooRTweet is the generalised successor to the CooRnet R package and implements the coordinated-link-sharing-behaviour (CLSB) methodology originally developed for Twitter analysis. It takes a social-media dataset (most cleanly Meta Content Library extracts, but flexible across platforms) and identifies clusters of accounts that share the same URLs (or other content tokens) within tight time windows, which is the canonical signature of an amplification operation. The package was published with peer review in Computational Communication Research 7(1) 2025 and is the most cited CIB methodology in academic communication research.

The "with CooRnet predecessor" framing matters operationally. CooRnet was the original R package, but its data ingest depended on CrowdTangle, which Meta deprecated in August 2024. CooRTweet was written as the generalised successor that decouples the methodology from any single data source: it works on any post-level dataset the analyst can shape into the package's expected schema. For SEA newsrooms with IFCN signatory status (Mafindo, Rappler, VERA Files, selected partners), Meta Content Library is the natural data source for CooRTweet runs because the IFCN gate is also the access path to Meta historical data; for newsrooms outside that gate, the package runs against any post-level extract the analyst can assemble.

When to use it

  • Your operation is an IFCN-signatory partner with Meta Content Library access and you want to run coordinated-link-sharing analysis on Facebook or Instagram historical data.
  • You are working in an R environment and need a peer-reviewed CLSB methodology that is the academic-community standard for publication-grade CIB findings.
  • An election-cycle investigation in Indonesia or the Philippines needs coordinated-amplification analysis on a multi-month dataset where temporal clustering of URL or content shares is the primary signal.
  • You are training researchers in a regional university partnership and want a tool whose underlying methodology is the most-cited reference in the academic CIB literature.

Limitations

Limitations

  • Requires R proficiency. The package is not a no-code tool; an analyst unfamiliar with R will spend significant ramp-up time before producing usable output.
  • Meta Content Library access is gated to IFCN-signatory research and fact-checking organisations. Mafindo, Rappler, and VERA Files have access; Laos has no IFCN signatory and therefore no MCL access path. Other SEA newsrooms vary.
  • CooRnet (the predecessor) depended on CrowdTangle, which Meta deprecated in August 2024. Older tutorials referencing CooRnet may assume CrowdTangle data flows that no longer exist; transition to CooRTweet was explicitly designed to remove this dependency.
  • Coordinated-link-sharing analysis surfaces necessary but not sufficient evidence of inauthenticity. Authentic political communities also share the same links in tight time windows; the analyst's interpretation is part of the method.

Privacy and threat model

CooRTweet runs locally on the analyst's machine within an R environment. The dataset stays on local infrastructure during analysis. The privacy concentration shifts to the input data source: Meta Content Library has its own data-handling and disclosure constraints attached to the IFCN access agreement, and any analysis output that names accounts is bound by those constraints.

For surveillance-environment work in Sri Lanka or Laos, the local- execution model is appropriate but the IFCN gate often blocks the underlying MCL data source. Where MCL access is unavailable, an analyst can substitute any post-level dataset they can assemble through other means (Telegram exports, custom scrapes for fringe platforms, archived Twitter API data where still accessible) at the cost of losing MCL's coverage of the largest SEA-region commercial platforms.

Country and platform applicability

  • Indonesia: language-agnostic; Mafindo's IFCN signatory status gives access to the MCL data path that CooRTweet handles natively. Relevant for buzzer-network and election-cycle analysis.
  • Laos: no IFCN signatory; no MCL access path; the package can run on any other post-level dataset the analyst assembles, at the cost of losing the dominant Lao-relevant Facebook data source. Honest gap.
  • Malaysia: Bernama MyCheck.My / Sebenarnya researchers with IFCN-equivalent access can run CooRTweet on MCL data.
  • Philippines: Rappler and VERA Files have IFCN signatory status; the package is operationally available to the #FactsFirstPH research layer.
  • Sri Lanka: FactCheck.lk's IFCN status was reported expired (regional case documentation Lanka); MCL access status should be reverified before relying on CooRTweet for SL work. Where MCL is unavailable, the package still runs on Telegram and other non-Meta extracts.
  • Thailand: AFP Fact Check Thailand's regional access via the AFP Medialab partnership may include MCL data flows; verify before relying on the path.

Platform applicability: native fit with Meta Content Library data (Facebook, Facebook Groups). Works on Telegram, historical Twitter extracts (X where still accessible), and any other post-level extract the analyst can shape to the package's schema.

How to access

Free download from the CRAN R repository or the package's GitHub mirror. Install via install.packages("CooRTweet") inside an R session, or from source via devtools::install_github() per the README. Meta Content Library access is the data-source gate, not the package itself; the IFCN signatory application path is a separate Meta workflow that institutional partners apply for independently.

Cost (current as of 2026-05)

Free. Open-source. The package itself imposes no cost. The Meta Content Library access path that the package handles natively is also free for IFCN-signatory partners but is gated by IFCN status rather than priced. The package is free in software terms but the most useful data source behind it is free-but-gated; for SEA newsrooms outside the IFCN gate (notably anywhere in Laos), the package's cost is functionally the cost of substitute data assembly.

Quickstart

  1. Install R 4.x and an IDE such as RStudio.
  2. Install the package: install.packages("CooRTweet").
  3. If Meta Content Library access is available, follow the package documentation to ingest MCL post-level data; otherwise, prepare a post-level extract from your alternative source following the package's expected schema.
  4. Choose the coordination construct: shared-URL coordinated link sharing is the canonical CooRTweet test, but the package supports other behavioural similarity constructs.
  5. Set the time window (typical: 60 seconds for very tight coordination, several minutes for amplification networks).
  6. Run the coordination-detection function and inspect the output network object.
  7. Export the network for visualisation in Gephi or for further analysis in Coordination Network Toolkit's similarity construct.
  8. Document the methodology and parameters in any published finding; cite the Righetti and Giglietto Computational Communication Research 7(1) 2025 paper for the underlying methodology.

In the toolkit's workflow

Behaviour-pillar non-detector tool per Architectural Anchor 1. Sits in 1C.1 Institutional CIB detection as the R-native alternative alongside CIB Mango Tree (Python interactive) and Coordination Network Toolkit (R or Python framework). The three together form a complementary open-source ladder; CooRTweet is the preferred path when the analyst is in R and the data is MCL.

Standard combinations:

  • With Meta Content Library at 2C.2 as the IFCN-gated data source.
  • With CIB Mango Tree for triangulation on the same dataset.
  • With Gephi at 2C.2 for publication-grade visualisation of CooRTweet network outputs.
  • With Yudistira (MAFINDO) at 2B.2 for Bahasa-specific claim extraction on the posts within identified clusters.
  • With ABCDE Framework and DISARM at 2C.3 for codifying the behavioural pattern in respectively public and institutional reporting.

This card carries no detector signal class: CooRTweet produces a non-detector behavioural-pattern signal class. Per Anchor 2, that signal combines with claim-extraction (2B.2) or source-history (1B.1) to support a publishable institutional finding.

Override notes

Override notes

  • Meta Content Library gating pointer (meta-content-library- gating-pointer): the package's most useful data path is gated to IFCN-signatory partners. For SEA newsrooms outside the gate (notably any Lao operation), the package is still usable on alternative datasets, but the bulk of Facebook-side coordination evidence is functionally inaccessible. Cost section above states this concretely.

Sources