Pith. sign in

REVIEW 4 major objections 5 minor 22 references

Creating a Historical Migration Dataset from Finnish Church Records, 1800-1920

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that a fully automated deep-learning pipeline can convert roughly 200,000 handwritten Finnish church migration records into a structured dataset of over six million entries.

desk verdict A valuable new dataset and an unusually honest pipeline paper, but 'suitable for research' overreaches: the unmeasured record-linkage step, not the HTR model, is the bottleneck. read the letter →

arxiv 2506.07960 v3 pith:L6E3W4SD submitted 2025-06-09 cs.CV

Ari Vesalainen , Jenna Kanerva , Aida Nitsch , Kiia Korsu , Ilari Larkiola , Laura Ruotsalainen , Filip Ginter This is my paper
classification cs.CV
keywords handwrittentextrecognitiondocumentlayoutanalysishistoricalmigrationrecordsFinnishchurchdeeplearningpipelinedemographicdatasetplacenamenormalization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that handwritten archival tables can be converted into structured, research-ready data without per-page manual transcription. The authors build a deep learning pipeline that de-skews scanned pages, detects table structure, classifies cells, recognizes handwritten text, and infers the year of each record, then apply it to roughly 200,000 images of Finnish church moving records from 1800 to 1920. The result is an open dataset of over six million migration entries. A case study of one parish shows how the extracted data can quantify arrivals and departures and map migration destinations, while also revealing how much cleaning and place-name normalization the raw output requires.

What carries the argument

The pipeline itself is the central mechanism. De-skewing is posed as a six-keypoint pose detection task solved with YOLO in two stages, so each page of an opening is straightened independently. Table and cell detection use YOLO followed by a density-based clustering step to reconstruct missing cells; Mask R-CNN performs pixel-level line detection to define rows and columns; a cell-type classifier skips empty and repetition cells; and a transformer-based handwriting recognition model fine-tuned for historical Finnish and Swedish transcribes each cell. Year extraction uses a YOLO detector plus the same handwriting recognition model, with a language model correcting the recognized years by enforcing a coherent sequence across pages. Because these stages are modular, each can be improved or replaced independently.

What would settle it

Take a random sample of released rows, say 1,000, manually transcribe the corresponding source-page images, and compare field by field after applying the paper's cleaning steps; if the agreement on date, migration direction, and standardized place name falls below the level needed for the intended demographic analysis, the claim that the dataset is suitable for research fails.

Watch

Extended reading notes

Core claim

The central claim is that large-scale automated extraction of structured data from handwritten historical records is feasible. The pipeline reaches component-level accuracies that support this: table detection F1 of 97.0, row detection F1 of 95.5, column detection F1 of 97.1, and cell-level handwriting recognition with a character error rate of 0.19. Applied to the full archive, it produced roughly 6.2 million rows from about 200,000 images in four days of parallel computing. The authors show the output can support demographic analysis, but the Elimäki case study makes clear that turning raw rows into usable records requires substantial post-processing: after duplication removal, field filtering, and place-name standardization, 60% of that parish's 18,809 extracted rows were usable, and only 8% of extracted parish-name spellings matched known parishes exactly before normalization.

Load-bearing premise

The whole enterprise depends on the assumption that the handwriting recognition output, with roughly one character in five wrong, is accurate enough that cleaning and normalization can produce a dataset whose remaining errors do not distort the demographic conclusions drawn from it.

Editorial extensions

If this is right

  • Historians can study internal migration volumes, flows, and destination patterns across Finland for 1800–1920 without manual transcription of the source books.
  • Linked with digitized birth, death, and disease records, the dataset enables studies of how migration shaped the spread of infectious diseases in pre-industrial Finland.
  • The modular pipeline can be adapted to other handwritten tabular archives, such as censuses, tax registers, or military rolls, with retraining on their layouts.
  • The Elimäki case demonstrates a post-processing recipe — duplication removal, heuristic column realignment, language-model-based place-name normalization, and manual review — that can turn noisy extractions into a clean parish-level migration dataset.
  • The open release of both the pipeline and the dataset lets other researchers reproduce the extraction and build their own cleaning layers.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The 'suitable for research' claim should be read as conditionally true: at the current character error rate, any study that relies on raw text fields will inherit noise, and the usable fraction of rows will vary by parish and layout.
  • Because the case-study cleaning used manual review and language-model normalization, a fully automated national-scale version of that cleaning step does not yet exist; the released dataset may therefore be most reliable for aggregate counts and less reliable for individual-level linkage.
  • The year-sequence correction via a language model suggests a reusable trick for other historical series where a field is monotonic or sequential; the same idea could correct misrecognized dates or page numbers in other archives.
  • Since free-text and half-table records, about 12% of the images, were excluded from the pipeline, migration events recorded outside tabular layouts are missing; a complete picture of Finnish internal migration would need a complementary method for those pages.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper presents a fully automated deep-learning pipeline for extracting structured migration records from roughly 200,000 images of Finnish church moving records (1800–1920). The pipeline combines de-skewing, table and line detection, cell-type classification, a National Archives HTR model for text recognition, and an LLM-based year-sequence correction. The authors evaluate each component on held-out annotated data and report strong component-level results (e.g., table detection F1 97.0, row detection F1 95.5, column detection F1 97.1), but also report that only 52% of test images have the correct number of tables and columns, that overall character error rate is 0.19, and that an Elimäki case study retains only 60% of extracted rows after cleaning and manual review. The paper releases the extracted dataset on Zenodo and the pipeline and annotations on GitHub, with the central claim that large-scale automated extraction of structured data from handwritten historical records is feasible.

Significance. If the central claim were fully supported, this would be an important contribution to historical demography and digital humanities: a multi-million-row migration dataset covering two centuries, with open code and data, would enable studies of internal migration, urbanization, and disease spread at an unprecedented scale. The authors are transparent about many limitations, and the release of the pipeline, annotated data, and dataset is a valuable scholarly resource. However, the significance is tempered because the dataset-level claim of being 'suitable for research' is only partially supported by the evidence presented: component accuracies are high, but end-to-end errors in table structure, text recognition, and place-name linking are substantial, and the one end-to-end case study required manual review and still discarded 40% of the initial rows. The work is best interpreted as demonstrating a scalable pipeline that produces a large, noisy extraction requiring per-parish cleaning, rather than a ready-to-use research dataset, unless additional dataset-level validation is provided.

major comments (4)
  1. [Abstract and Section 6] The abstract states that applying the pipeline 'resulted in a structured dataset suitable for research,' and Section 6 concludes that the work 'demonstrates that large-scale, automated extraction of structured data from handwritten historical records is feasible.' These claims are not supported by the paper's own end-to-end numbers. Section 4 reports that in 48% of test images (92/192) the predicted number of tables or columns does not match the ground truth, and Section 5 shows that after cleaning and manual review only 60% of Elimäki rows remain usable. No end-to-end measure of full-record correctness (correct row segmentation, correct column alignment, correct cell text, correct year, and successful place linkage) is reported. The paper should either soften the claims to describe a large-scale pipeline producing a noisy dataset that requires cleaning, or provide dataset-level quality metrics that directly quantify the fraction of complete, research-usable records.
  2. [Section 3.7 and Section 4] The component evaluations in Tables 5–9 are performed on separate test sets and are informative, but they do not quantify the compounding of errors across stages. For example, a record is usable only if the row and column structure is correct, the HTR output for the date/name/parish is accurate, and the year is correctly assigned. Section 4 explicitly notes 48% image-level table/column errors and only approximately 91% row extraction excluding double-page splits. The paper never reports the accuracy of a complete extracted record (e.g., precision and recall of rows whose date, name, parish, and year are all correct). Without such an end-to-end metric, the 'suitable for research' claim rests on an unmeasured multiplicative error chain. I ask the authors to provide a record-level evaluation or to explicitly limit the claim to specific downstream tasks that tolerate character-level noise.
  3. [Section 5.1] The case study reveals a silent temporal-data loss that is not quantified for the full dataset: 'there are no records of arrivals to Elimäki for the years 1914 and 1915. This is due to the year identification failing in few pages, and the records from these years are merged with records from 1916 and 1917.' Since year is a critical variable for demographic analysis, the paper should either quantify how many pages across the full dataset suffer from this year-merge problem or otherwise demonstrate that such failures are rare outside Elimäki. As written, a user of the released dataset cannot know which years or pages are affected, which is a load-bearing uncertainty for the dataset's research usability.
  4. [Section 5, Table 11] Table 11 shows that only 8% of extracted parish names match a known parish at edit distance 0, and 23% at edit distance ≤1. The subsequent LLM-based cleaning plus manual review raises linkage to 66% in the Elimäki case study, but that manual review is not part of the automated pipeline and was not applied to the other 467 parishes. The paper's claim that the pipeline is 'automated extraction' is technically accurate, but the path from raw extraction to a research-ready dataset is not automated and its cost, accuracy, and generalizability are unmeasured. The paper should state clearly that automatic place-name linking is a major unsolved stage for the rest of the dataset, or provide an automatic linking method with quantitative evaluation.
minor comments (5)
  1. [Section 4] The 91% row extraction figure excludes the 16 double-page split images; it would be helpful to also report row extraction including those images, or to explain how double-page splits are handled in the released CSV files.
  2. [Section 3.7.4] The text recognition evaluation excludes 342 lines containing question marks (unreadable to human annotators); the paper should state how many lines in the full dataset are likely to be affected by this unreadability, as it may affect the reliability of the extracted text.
  3. [Section 3.6] The year-detection evaluation in Table 10 reports precision and recall for year mentions on annotated pages, but does not report the downstream per-record year accuracy after LLM sequence correction; a per-record or per-page-level year accuracy would be more directly relevant to users.
  4. [Section 5] The Elimäki case study notes that 64% of rows followed the expected layout and an additional 34% were realigned using heuristics; the paper does not describe these heuristics in enough detail for a reader to reproduce or assess them, and a description or reference would improve reproducibility.
  5. [Throughout] The manuscript contains several minor typographical and formatting issues, such as the inconsistent spacing in author names and the non-standard use of 'A ¨ ıda' and 'a ∗†'; I recommend a careful copyedit.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: pipeline components are evaluated on held-out annotations, and the Elimäki limitations are empirical, not derived from fitted parameters.

full rationale

I find no circular step in the derivation chain. The pipeline stages—de-skew, table and line detection, cell classification, text recognition, and year extraction—are trained on manually annotated data and evaluated on held-out test images with separate annotations (Tables 3–10); no evaluation quantity is computed from the fitted parameters themselves. The HTR model is an external National Archives model cited as Kansallisarkisto (2024), not a self-citation, and it is used without modification. The dataset-level feasibility claim is supported by component metrics and by the Elimäki case study, which the paper itself reports as lossy: only 8% of raw parish names match at edit distance zero, 60% of initial records remain usable, and year extraction fails for 1914–1915 in that parish. These are independent empirical limitations, not inputs that are later renamed as predictions. The only overlapping-author citations (Nitsch et al., 2016, 2023, 2025) are used for contextual comparison of migration rates and as motivation, not as load-bearing justification of the extraction result. Therefore no reduction by construction occurs; the score of 1 reflects minor non-load-bearing self-citations only.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The pipeline uses trained model weights, but those are learned parameters fitted to annotated data, not free parameters of a scientific derivation. The axioms listed are the domain assumptions that must hold for the dataset to have demographic meaning.

assumptions (3)
  • domain assumption Church migration records are an accurate and complete record of internal migration in Finland during 1800-1920.
    The dataset's demographic value depends on this; introduced in Section 2.
  • domain assumption Rows in the extracted tables correspond to individual migration events or family moves.
    Used in Section 4 and 5 to count departures and arrivals.
  • domain assumption The manually annotated ground truth is correct and representative of the full collection.
    All component evaluations rely on the 1,632 annotated images described in Section 2.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Creating a Historical Migration Dataset from Finnish Church Records, 1800-1920." pith.science (2026). https://pith.science/paper/L6E3W4SD

@misc{pith2026250607960,
  author       = {Pith},
  title        = {Pith review of: Creating a Historical Migration Dataset from Finnish Church Records, 1800-1920},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6E3W4SD}},
  note         = {Machine review of arXiv:2506.07960}
}
read the original abstract

This article presents a large-scale effort to create a structured dataset of internal migration in Finland between 1800 and 1920 using digitized church moving records. These records, maintained by Evangelical-Lutheran parishes, document the migration of individuals and families and offer a valuable source for studying historical demographic patterns. The dataset includes over six million entries extracted from approximately 200,000 images of handwritten migration records. The data extraction process was automated using a deep learning pipeline that included layout analysis, table detection, cell classification, and handwriting recognition. The complete pipeline was applied to all images, resulting in a structured dataset suitable for research. The dataset can be used to study internal migration, urbanization, and family migration, and the spread of disease in preindustrial Finland. A case study from the Elim\"aki parish shows how local migration histories can be reconstructed. The work demonstrates how large volumes of handwritten archival material can be transformed into structured data to support historical and demographic research.

Figures

Figures reproduced from arXiv: 2506.07960 by the authors.

Figure 1
Figure 1. Cumulative counts for different preprinted layout types. Typically, in- and out-migrations were recorded in separate books, or in separate tables within a single book, with one table per page of a book opening. Occasionally, these were combined into a single table with separate columns for in- and out-migrants. Some variations in table structure occurred depending on regional practices or the instructions followed b… view at source ↗
Figure 2
Figure 2. Example of handdrawn moving table (Huittinen 1878). FFHA’s digital archive. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Example of preprinted moving table (Hein¨avesi 1909). FFHA’s digital archive. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Details of a typical moving table entry from Hankasalmi: Maria Sirkka, a servant (piika), moved [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Text recognition for tabular data. We approach de-skew as an image recognition problem, where the objective is to recognize six key points on the image: the four corners of the opening, and the two ends of the middle division. The idealized process is illustrated in […
Figure 6
Figure 6. Figure 6: Extreme example of page skew (left) and the output of the de-skew process (right). Red circles [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: De-skew process. Two pages of the opening with the relevant six keypoints A-F and the image [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Example of how clustering improves results. In some cases, the table cell detection model fails [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Example of an opening with several year mentions outside of the header area. [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: Histograms of departures from and arrivals to Elim¨aki between 1875 and 1922. [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]
Figure 11
Figure 11. Figure 11: Maps showing the origins and destinations of migration to and from Elim¨aki between 1875 [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

22 extracted references · 18 canonical work pages

  1. [1]

    write newline

    " write newline " cite write " FUNCTION editor.postfix editor num.names #1 > "( )" "( )" if FUNCTION editor.trans.postfix editor num.names #1 > "( )" "( )" if FUNCTION trans.postfix translator num.names #1 > "( )" "( )" if FUNCTION authors.editors.reflist.apa5 'field := 'dot := field num.names 'numnames := numnames 'format.num.names := format.num.names na...

  2. [2]

    APACrefauthors \ 1953

    sukututkimusaineiston1953 APACrefauthors Anonymous. APACrefauthors \ 1953 . Sukututkimusaineiston valokuvaaminen Sukututkimusaineiston valokuvaaminen . Genos 3 89 . In Finnish

    work page 1953

  3. [3]

    , Enflo, K

    Blomqvist2022Joint APACrefauthors Blomqvist, C. , Enflo, K. , Jakobsson, A. \ Åström, K. APACrefauthors \ 2022 . Joint Handwritten Text Recognition and Word Classification for Tabular Information Extraction Joint handwritten text recognition and word classification for tabular information extraction . 26th International Conference on Pattern Recognition (...

    arXiv 2022

  4. [4]

    , Ketola, T

    Briga2022The APACrefauthors Briga, M. , Ketola, T. \ Lummaa, V. APACrefauthors \ 2022 . The epidemic dynamics of three childhood infections and the impact of first vaccination in 18th and 19th century F inland The epidemic dynamics of three childhood infections and the impact of first vaccination in 18th and 19th century F inland . medRxiv . doi:10.1101/2...

    work page doi:10.1101/2022.10.30.22281707 2022

  5. [5]

    , D \'e jean, H

    Clinchant2018Comparing APACrefauthors Clinchant, S. , D \'e jean, H. , Meunier, J L. , Lang, E M. \ Kleber, F. APACrefauthors \ 2018 . Comparing machine learning approaches for table recognition in historical register books Comparing machine learning approaches for table recognition in historical register books . 13th IAPR International Workshop on Docume...

    work page 2018

  6. [6]

    , Kahle, P

    9041761 APACrefauthors Colutto, S. , Kahle, P. , Guenter, H. \ Muehlberger, G. APACrefauthors \ 2019 . Transkribus. A Platform for Automated Text Recognition and Searching of Historical Documents Transkribus. a platform for automated text recognition and searching of historical documents . 15th International Conference on eScience (eScience) 15th internat...

    arXiv 2019

  7. [7]

    , Ibrayim, M

    deng2024yolo APACrefauthors Deng, Q. , Ibrayim, M. , Hamdulla, A. \ Zhang, C. APACrefauthors \ 2024 . The YOLO model that still excels in document layout analysis The YOLO model that still excels in document layout analysis . Signal, Image and Video Processing 18 2 1539--1548

    work page 2024

  8. [8]

    APACrefauthors \ 2025

    sshy_website APACrefauthors Finland's Family History Association (FFHA) . APACrefauthors \ 2025 . Finland's Family History Association Website. Finland's family history association website. https://www.sukuhistoria.fi/sshy/index.htm . Accessed: January 8, 2025

    work page 2025

Show all 22 references
  1. [9]

    , Khoroshiltseva, M

    FIORUCCI2020102 APACrefauthors Fiorucci, M. , Khoroshiltseva, M. , Pontil, M. , Traviglia, A. , Del Bue , A. \ James, S. APACrefauthors \ 2020 . Machine Learning for Cultural Heritage: A Survey Machine learning for cultural heritage: A survey . Pattern Recognition Letters 133 ...

    2020 doi

  2. [10]

    , Romero, V

    Granell2023Processing APACrefauthors Granell, E. , Romero, V. , Prieto, J R. , Andrés, J. , Quirós, L. , Sánchez, J A. \ Vidal, E. APACrefauthors \ 2023 . Processing a large collection of historical tabular images Processing a large collection of historical tabular images . Pa...

    2023 doi

  3. [11]

    , Gkioxari, G

    he2017mask APACrefauthors He, K. , Gkioxari, G. , Doll \'a r, P. \ Girshick, R. APACrefauthors \ 2017 . Mask r-cnn Mask r-cnn . Proceedings of the IEEE international conference on computer vision Proceedings of the IEEE international conference on computer vision \ ( \ 2961--2969)

    2017

  4. [12]

    APACrefauthors \ 1981

    Hietala1981Internal APACrefauthors Hietala, K. APACrefauthors \ 1981 . Internal migration and technological development Internal migration and technological development . Finnish Yearbook of Population Research . doi:10.23979/fypr.44751

    1981 doi

  5. [13]

    , Yan, Q

    huang2019yolo APACrefauthors Huang, Y. , Yan, Q. , Li, Y. , Chen, Y. , Wang, X. , Gao, L. \ Tang, Z. APACrefauthors \ 2019 . A YOLO -based table detection method A YOLO -based table detection method . Proceedings of the 15th International Conference on Document Analysis and Re...

    2019

  6. [14]

    , Briga, M

    ketola2021Town APACrefauthors Ketola, T. , Briga, M. , Honkola, T. \ Lummaa, V. APACrefauthors \ 2021 . Town population size and structuring into villages and households drive infectious disease risks in pre-healthcare F inland Town population size and structuring into village...

    2021

  7. [15]

    , Burghardt, M

    Lehenmeier2020Layout APACrefauthors Lehenmeier, C. , Burghardt, M. \ Mischka, B. APACrefauthors \ 2020 . Layout Detection and Table Recognition – Recent Challenges in Digitizing Historical Documents and Handwritten Tabular Data Layout detection and table recognition – recent c...

    2020

  8. [16]

    li2023trocr APACrefauthors Li, M. , Lv, T. , Chen, J. , Cui, L. , Lu, Y. , Florencio, D. Wei, F. APACrefauthors \ 2023 . TrOCR : Transformer-based optical character recognition with pre-trained models TrOCR : Transformer-based optical character recognition with pre-trained mod...

    2023

  9. [17]

    , Lummaa, V

    nitsch2025The APACrefauthors Nitsch, A. , Lummaa, V. , Ketola, T. , Honkola, T. , Vesakoski, O. \ Briga, M. APACrefauthors \ 2025 . The spatial distribution of pertussis, but not measles or smallpox, in pre-industrial F inland matches dialects The spatial distribution of pertu...

    2025

  10. [18]

    , Gooding, P

    nockels2022understanding APACrefauthors Nockels, J. , Gooding, P. , Ames, S. \ Terras, M. APACrefauthors \ 2022 . Understanding the application of handwritten text recognition technology in heritage contexts: a systematic review of T ranskribus in published research Understand...

    2022

  11. [19]

    , Helske, J

    pasanen2024spatio APACrefauthors Pasanen, T M. , Helske, J. , H \"o gmander, H. \ Ketola, T. APACrefauthors \ 2024 . Spatio-temporal modeling of co-dynamics of smallpox, measles, and pertussis in pre-healthcare F inland Spatio-temporal modeling of co-dynamics of smallpox, meas...

    2024

  12. [20]

    APACrefauthors \ 1980

    pitkanen1980registering APACrefauthors Pitk \"a nen, K. APACrefauthors \ 1980 . Registering people in a changing society-the case of F inland Registering people in a changing society-the case of F inland . Finnish Yearbook of Population Research 60--79

    1980

  13. [21]

    , Malyuk, M

    LabelStudio APACrefauthors Tkachenko, M. , Malyuk, M. , Holmanyuk, A. \ Liubimov, N. APACrefauthors \ 2020-2025 . Label Studio : Data labeling software. Label Studio : Data labeling software. APACrefURL https://github.com/HumanSignal/label-studio APACrefURL Open source softwar...

    2020

  14. [22]

    APACrefauthors \ 2024

    ultralytics2023yolov11 APACrefauthors Ultralytics. APACrefauthors \ 2024 . YOLOv11 implementation. Yolov11 implementation. https://github.com/ultralytics/ ultralytics . GitHub . Retrieved from GitHub on 2024-12-01

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.