birddog

The goal of birddog is sniffing out emergence and trajectories in scientific and patent literature.

Installation

Install the stable version from CRAN:

install.packages("birddog")
library(birddog)

Or the development version from GitHub:

# install.packages("remotes")
# remotes::install_github("roneyfraga/birddog")
install.packages("birddog", repos = c('https://roneyfraga.r-universe.dev', 'https://cloud.r-project.org'))
library(birddog)

Features

Data import

read_openalex() – OpenAlex API or CSV exports
read_wos() – Web of Science BibTeX, RIS, plain-text, tab-delimited

Citation network and community detection

sniff_network() – direct citation or bibliographic coupling networks
sniff_components() – identify connected components
sniff_groups() – community detection (fast greedy, Louvain, Leiden, walktrap, edge betweenness)

Group analysis

sniff_groups_attributes() – group-level summary statistics and horizon plots
sniff_groups_keywords() – keyword frequency per group
sniff_groups_terms() – NLP-based phrase extraction
sniff_groups_hubs() – hub classification (Zi-Pi, Guimera and Amaral 2005)
sniff_groups_cumulative_citations() – per-document citation growth
sniff_groups_influence() – directed citation influence between groups: cross-citation matrix, debt / audience / Salton / surprise indices, net flow, and source / broker / sink roles
plot_groups_influence_matrix() – directed influence heatmap, the asymmetric twin of the confluence matrix
plot_groups_influence_network() – net-influence spine (who, on balance, leads whom)

Indexes

sniff_cct() – measures the pace of change (Kayal 1999)
sniff_entropy() – normalized Shannon entropy for keyword diversity (Shannon 1948; Pielou 1966)

Trajectory detection

sniff_groups_cumulative() – cumulative clusterization over time
sniff_groups_lineage() – Jaccard similarity DAG across years
plot_groups_lineage_2d() / plot_groups_lineage_3d() – node-based lineage plots
sniff_trajectory_dag() / plot_trajectory_dag() – temporal DAG of cluster-year nodes across the whole history (plot_trajectory_dag_interactive() for the plotly view); similarity = "coupling" routes the DAG by shared references instead of document overlap
sniff_trajectory_braid() – decompose the DAG into trajectories: one central per final group (tr::cNgN) plus the absorbed tributaries that merge into it (trN)
sniff_trajectory_channel() – sibling detector that routes one global optimal-path backbone per final group (potential-routed), an alternative decomposition of the same DAG
subset() – focus a flow on a watershed or a predicate while it stays a valid flow (subset(flow, target = "tr::c1g1"))
is_flow() / validate_flow() – test and validate the birddog_flow object contract
sniff_trajectory_coherence() – score how content-coherent a flow’s partition is, with a silhouette over an independent content signal
sniff_trajectory_comparison() – compare two detectors head-to-head by content coherence, reporting the contested nodes

Trajectory analysis

sniff_trajectory_dynamics() – per-trajectory dynamic-state indicators and life-cycle classification (emergence index, staying power, reach), with optional CCT, entropy, hubs and community lenses joined in
sniff_trajectory_cct() – per-year citation cycle time (renewal pace) along each trajectory
sniff_trajectory_entropy() – per-year keyword diversity (Pielou’s J’) along each trajectory
sniff_trajectory_hubs() – hub roles aggregated to each trajectory (provincial vs bridging)
sniff_trajectory_community() – community breadth (distinct authors) of each trajectory
sniff_trajectory_emergence_owners() – the authors who own each living trajectory’s emergence
sniff_trajectory_self_sufficiency() – per-trajectory endogenous-growth index (1 - imported / size)
sniff_trajectory_destination() – track where a dying trajectory’s papers go, naming the trajectory that absorbed it
sniff_trajectory_formation() – the inverse of destination: which trajectories fed into a target (its feeders)
sniff_trajectory_contribution() – documents an intermediate trajectory contributes to a target in a given year
sniff_trajectory_confluence() – render-ready confluence forest of the soft DAG (how the central trajectories form)

Trajectory plots

plot_trajectory_dynamics() / plot_trajectory_dynamics_interactive() – strategic map of trajectory dynamic states
plot_trajectory_confluence() / plot_trajectory_confluence_interactive() / plot_trajectory_confluence_matrix() – braided-river confluence of the central trajectories
plot_trajectory_formation() – confluence of feeders into a target trajectory (cumulative river timeline)
plot_trajectory_dispersion() – timeline of the stagnant trajectory and the one that absorbed it
plot_trajectory_lines_2d() / plot_trajectory_lines_3d() – variable-width trajectory line plots on a time layout

Main path analysis

sniff_key_route() – key-route search (Liu and Lu 2012) with SPC weights

Topic modeling

sniff_groups_stm_prepare() / sniff_groups_stm_run() – structural topic modeling within groups

Vignettes

The vignettes are available online here:

Methodological workflow

Main publications

Miranda et al. (2025) The Landscape of Green and Biohydrogen Technology: A Data-Driven Exploration Using Non-Supervised Methods
Felizardo et al. (2025) Transforming Wastes into Resources: Innovations in Cotton Biorefineries for a Sustainable Future
Biazatti et al. (2024) Soybean biorefinery and technological forecasts based on a bibliometric analysis and network mapping
Maria et al. (2023) Evolution of Green Finance: A Bibliometric Analysis through Complex Networks and Machine Learning
Matos et al. (2023) Building and evaluating prospective scenarios for corn-based biorefineries
Souza et al. (2022) Is entrepreneurship an emerging area of research? A computational response
Souza et al. (2022) Bioenergy research in Brazil: A bibliometric evaluation of the BIOEN Research Program