Screening data and code for a systematic review of technological tools in rice farming (2016–2026)

1. Purpose: Systematic reviews are reproducible only if the study-selection process can be independently re-executed. This deposit makes that possible. Any researcher who re-runs the search query in Scopus and applies the supplied script will obtain the identical eligible corpus (n = 13,106) and the identical cluster prevalence figures reported in the article. 2. Search query, Executed in Scopus Advanced Search on 5 August 2026: ``` TITLE-ABS-KEY ( ( rice OR paddy ) AND ( "internet of things" OR iot OR sensor* OR drone* OR uav OR "unmanned aerial" OR "remote sensing" OR gis OR gps OR gnss OR "machine learning" OR "artificial intelligence" OR "deep learning" OR "big data" OR "digital twin" OR robot* OR "agricultural machinery" OR mechaniz* OR "decision support system" ) ) ``` Applied limits: `PUBYEAR > 2015`; `LANGUAGE = English`; `DOCTYPE = ar, cp, ch`. Export settings: all citation information, abstract and keywords, and affiliation fields selected. The export returned 20,000 records, which is the Scopus export ceiling; this constraint is disclosed as a limitation in the article. 3. Contents ``` README.md This file CODEBOOK.md Variable definitions and cluster coding rules scripts/ screening_pipeline.py Full PRISMA screening and cluster coding data/ eligible_corpus_metadata.csv Derived corpus, n = 13,106 (see §5) cluster_prevalence.csv Cluster counts and shares (reproduces Table 4) prisma_counts.json Stage-by-stage exclusion counts (reproduces Table 2) ```