Input and output
1,686 pages source pages to 16,001 biographies linked, structured records in the finished release.Degener Dataset
Reproducible LLM workflows in economic history
Degener's
Wer Ist's?
A transparent pipeline for turning Degener's Wer Ist's? (1911) into a structured dataset with prompts, logs, validation steps, and reusable outputs.
The website accompanies my paper on reproducible historical data construction. For details, see below or inspect the workflow.
Structured Data
Biography record
- Name
- WEBER
- First names
- Max
- Titles/profession
- Dr. jur., o. Hon.-Prof. Nat.-Ökon. Univ. Heidelberg.
- Gender
- male
- Birth date
- 21. April 1864
- Birth place
- Erfurt.
- Address
- Heidelberg, Ziegelhäuser Landstraße 17.
- Education
- Königliches Gymnasium Charlottenburg; Universität Heidelberg, Straßburg, Berlin, Göttingen; Schüler von L. Goldschmidt und A. Meitzen; philosophisch von H. Rickert beeinflusst; Doktor der Rechte 89.
- Job / occupation
- Dr. jur., o. Hon.-Prof. Nat.-Ökon. Univ. Heidelberg.
- Career
- Refer. 86; Dr. jur. 89; Assess. Berlin 90; Priv.-Doz. 92; a. o. Prof. Eidebrecht. Berlin 93; o. Prof. Nat.-Ök. Freiburg 94; Heidelberg 97; künd. Lehramt weg. hartnäck. Krankht. 03; sd. o. Hon.-Prof. das.
- Father
- Max W., Stadtr. Berlin, Rchstgsabgeordn.
- Mother
- Helene Fallenstein.
- Ancestors
- Vorfahren mütterlicherseits: Hugenotten.
- Spouse
- Verheiratet: mit Marianne Schnitger, Tochter des Sanitätsrats Doktor S., Lippe, bekannte Schriftstellerin auf dem Gebiet der Philosophie und Rechtsgeschichte der Ehe.
- Children
- unknown
- Works/publications
- Z. Gesch. d. Hdeisges. 89; Röm. Agrargesch. 92; Ostelb. Ldarbeiter. 93; Protestant. Ethik u. Geist d. Kapitalismus (Arch. f. Sozialwiss., Bd. XXI.); zahlr. methodol. u. Psychophysik d. gewerbl. Arb. betr. Abhdlgn. das.; Mithrsg. d. Arch. f. Soz.-Wissensch.
- Specialization
- Spez.: Rechtsgeschichte; Nationalökonomie; Soziologie; Methodologie.
- Hobbies
- unknown
- Collections
- unknown
- Personal notes
- unknown
- Political party
- unknown
- Memberships
- außerordentliches Mitglied Akademie der Wissenschaften Heidelberg.
Featured Paper
Working paper
Reproducible LLM Workflows in Economic History: Degener's Wer Ist's?
The paper develops a transparent, reproducible LLM workflow for turning dense historical print sources into structured data. Degener's Wer Ist's? (1911) is the demonstration case: roughly 16,000 biographies are processed through OCR, assembly, extraction, validation, and normalization.
The central argument is practical: LLM-based data construction becomes more useful when prompts, logs, model responses, and outputs remain linked to the original source.
Main results
Variables
22 Normalized fields capture names, occupations, addresses, family details, publications, and source-quality flags.Quality check
4.49% semantic error rate from human verification. 1.07% CER sample estimate.Efficiency
€159.24 total model cost. 66 min total processing time.