% 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) & 15{,}490{,}159 \\
Excluded E1 outside the window & 160 \\
Excluded E2 non-research document type & 3{,}961{,}589 \\
Excluded E3 XPAC expansion record (not DBLP-linked) & 1{,}421{,}692 \\
Excluded E4 no title & 1{,}188 \\
Excluded E5 preprint merged into its published version & 106{,}928 \\
After exclusions & 9{,}998{,}602 \\
CS core (mapping corpus) & 6{,}403{,}238 \\
CS penumbra (sensitivity set) & 3{,}595{,}364 \\
Strict core: DBLP-indexed and primary field CS & 1{,}868{,}188 \\
\bottomrule
\end{tabular}
\end{threeparttable}\end{table}