Corpus charts
The structure of the Robert-database ingestion — what the 55,749 enriched extracted tables look like in aggregate. Each chart is a genuine count from the Robert database (no constructed economic series — these measure the corpus), with a top-right CSV/Parquet download of the exact plotted data.
Two Vladimirs, one corpus
The corpus records a century and a half of official statistical measurement across two Russian regimes — the lens of the two Vladimirs the site is named for. Lenin's material, the structure of production: industry, agriculture, labour, province by province. Voytinsky's material, the comparable aggregate: money, budgets, prices and trade, assembled into series that can be set beside one another. Both are densely represented below — industry (4,220 tables) and finance & budget (3,017) are the second and third largest named topics after foreign trade — and the decade histogram shows the measurement effort surviving war, revolution and reform.
Read the charts with this caveat. The topic and geography charts describe the 25,677 tables that have been classified against the ratified taxonomy, not all 55,749 in the catalog. The remaining 30,072 were metadata-enriched later and have not yet been through the classification pass; they appear honestly as (unclassified) — the single largest bar — rather than being dropped from the total or assigned a guessed topic. The same holds for the decade histogram, which can only place the 25,313 tables whose period was recovered. Shares below are therefore shares of the classified subset.
Enriched tables by topic
The ratified 16-category USSR statistical taxonomy. (unclassified) is the largest bar — 30,072 tables awaiting the classification pass, shown rather than hidden. Among the classified, foreign trade dominates (the largest single extraction), with industry and finance/budget next — the planning and monetary record at the heart of the Soviet statistical state.
Enriched tables by decade
When the measured years fall (parsed from each table's period coverage). The imperial fiscal/military record (1860s–1910s), the early-Soviet rebuild, and the dense late-Soviet statistical yearbooks (1940s–1980s) are all visible.
Enriched tables by geography
The geographic level each table covers (USSR total, union republic, oblast/krai, RSFSR, economic region, imperial Russia, or foreign/other) — the facet that makes the corpus a candidate regional panel.
Transcription quality (two-axis model)
The Robert framework records two orthogonal quality axes. This chart shows the transcription-status distribution — how the cell values were obtained: V verified, H human-grade, L LLM-read, R needs repair, X failed. Most tables are LLM-read (honest about the extraction method); a verified/human-grade core anchors the quality. Nothing is over-claimed.