% generated by tables/make_tables.py from data/metadata/prisma_counts.json
\begin{table}[htbp]\centering\caption{Record flow of the mapping corpus (PRISMA-style, sequential exclusions).}\label{tab:prisma}
\begin{threeparttable}\small
\begin{tabular}{lr}
\toprule
\textbf{Stage} & \textbf{Records} \\
\midrule
Identified: OpenAlex works with a CS topic, 2018–2026 (snapshot 2026-10-14, both passes) & 16{,}592{,}966 \\
Excluded E1 outside the window & 59{,}453 \\
Excluded E2 non-research document type & 3{,}994{,}472 \\
Excluded E3 XPAC expansion record (not DBLP-linked) & 1{,}430{,}851 \\
Excluded E4 no title & 1{,}193 \\
Excluded E5 preprint merged into its published version & 114{,}421 \\
Excluded E7 generic deposit platform without DBLP/arXiv support & 473{,}549 \\
After exclusions & 10{,}519{,}027 \\
CS core (mapping corpus) & 7{,}110{,}289 \\
CS penumbra (sensitivity set) & 3{,}402{,}475 \\
Strict core: DBLP-indexed and primary field CS & 1{,}832{,}623 \\
\bottomrule
\end{tabular}
\end{threeparttable}\end{table}