NEWS

SportMiner 0.1.0 (2026-01-17)

Initial CRAN Release

New Features

API Functions
- sm_set_api_key(): Secure API key management via environment variables
- sm_search_scopus(): Retrieve papers from Scopus with automatic pagination
- sm_get_indexed_keywords(): Extract indexed keywords for individual papers
Text Processing
- sm_preprocess_text(): Tokenize, remove stopwords, and stem abstracts
- sm_create_dtm(): Generate document-term matrices with flexible filtering
Topic Modeling
- sm_select_optimal_k(): Coherence-based selection of topic count
- sm_train_lda(): Train Latent Dirichlet Allocation models
- sm_compare_models(): Multi-model comparison (LDA, STM, CTM) with metrics
Visualizations
- sm_plot_topic_terms(): Bar charts of top terms per topic
- sm_plot_topic_frequency(): Document distribution across topics
- sm_plot_topic_trends(): Stacked percentage charts showing temporal trends
- theme_sportminer(): Custom colorblind-friendly ggplot2 theme
Network Analysis
- sm_keyword_network(): Co-occurrence network visualization

CRAN Compliance

All API calls wrapped in tryCatch() for graceful error handling
Global variables properly declared using .data pronoun from rlang
Message/warning functions used instead of cat() or print()
Comprehensive test suite with offline-safe mocks via httptest
Zero errors, warnings, or notes in R CMD check --as-cran

Documentation

Complete roxygen2 documentation for all exported functions
"Getting Started" vignette with end-to-end workflow example
README with quick start guide and advanced usage examples

Testing

30+ unit tests covering all major functions
Edge case handling (empty data, missing columns, invalid parameters)
Mock API responses for offline testing
Tests marked with skip_on_cran() for computationally expensive operations

Note: This is the initial release. Future versions will include additional topic modeling algorithms, enhanced network analysis features, and extended visualization options.