ADISP1936SUSUnavailable

Paris Population Censuses - 1926-1936

Machine translation

  • Base POPP (Projet d’Océrisation des Recensements de la Population Parisienne) (1926-1936)

Some metadata is not available in this language. The original text is shown and flagged.

Summary

The POPP project aimed to develop a large database (of almost 9 million people) based on the nominative lists of censuses in Paris of 1926, 1931 and 1936, the only ones existing for the Parisian population (with that of 1946). The POPP database contains all the information contained in the nominal lists of the population (address, surname, given names, date and place of birth, nationality, marital status, situation in the household, occupation), to which various variables have been added, including the variable "sex" (not included in the nominal lists), a variable accounting for absent persons or a typology of occupations created by matching the professions declared by Parisians to the professional classification of the 1931 census. POPP is the first French demography base created thanks to the progress of Artificial Intelligence and, in particular, Deep Learning. The POPP project is thus a multidisciplinary project between a team of computer scientists and a team of SHS researchers.

Machine translation

French title

Recensements de la Population Parisienne - 1926-1936

Topics

Keywords

  • Population
  • Demographic statistics
  • Demography
  • War

Bibliographic citation

doi:10.13144/lil-1719 Recensements de la Population Parisienne - 1926-1936 - CNRS (producteur), Brée, Sandra ; Paquet, Thierry ; Constum, Thomas ; Tranouez, Pierrick ; Gay, Victor ; Kempf, Nicolas ; Leturcq, Marion (auteurs), PROGEDO-ADISP (diffuseur).

Record production date

1936

Producers

  • Centre National de la Recherche Scientifique (CNRS)

Other identifications / acknowledgements

  • Buidze, Mariia — Membre de l’équipe de curation et d’adaptation de la base à la recherche en SHSCNRS-ObHisPop
  • Doignon, Yoann — Membre de l’équipe de curation et d’adaptation de la base à la recherche en SHSCNRS, Idées
  • Coulmont, Baptiste — Membre de l’équipe de curation et d’adaptation de la base à la recherche en SHSENS Paris-Saclay
  • Sidorets, Anna — Aide au nettoyage des donnéesCNRS-ObHisPop
  • Merveille, François — Co-porteur du projet POPPGED-Campus Condorcet

Distributor

Record version

Version 1 (07/07/2025): documentation of the upper part and publication of the V0 (distribution pending before receipt of final data).

Version date : Jul 7, 2025

Details

Parallel titles

  • Paris Population Censuses - 1926-1936

Country

  • France (FR)

Geographic coverage

  • Paris

Geographic unit : Address (a Geographic Information System is being created for this database)

Analysis unit

  • Individual / Household

Data type

  • Individual census data

Study population

Parisian population recorded in censuses of 1926, 1931 and 1936

Collection dates

  • 2020 Start
  • 2024 End

Methodology

Time method of the data collection

  • Cross-section

Sampling procedure

  • Total universe/Complete enumeration

Frequency of the data collection : Every 5 years

Data collection method : Compilation / synthesis The dataset results from the following operations: - Use of the images of the nominative lists of the population censuses of Paris that were previously digitised by the Archives de Paris and are available on their site; with the exception of populations counted separately: https://archives.paris.fr/s/11/denombrements-de-population/? - Manual annotations of 243 double pages "ground truth" (223 for machine training; 40 for testing) -Analysis of document images by deep neural networks for the detection of table lines; HTR (Handwriting Text Recognition) and syntactic Recognition of Manuscript Names Using Weighted Finite-State Transducers (WFST) - The whole process of creating the "version 0" database (i.e. before curation, correction and adaptation) is detailed in the data paper: Constum, T. et al. (2022). Recognition and Information Extraction in Historical handwritten Tables: Toward Understanding Early 20th Century Paris Census. In: Uchida, S., Barney, E., Eglin, V. (eds) Document Analysis Systems. DAS 2022. Lecture Notes in Computer Science, vol. 13237. Springer, Cham. https://doi.org/10.1007/978-3-031-06555-2_10

Type of research instrument : Technical instrument(s) Programming script Data collection instructions

Methodology notes : Despite manual cleaning, it is possible that there are still errors, especially related to the column jump poorly understood by the machine. These errors are considered non-essential and non-disabled for statistical analysis of data.

Processings

Cleaning operations : All the curations, adaptations, creations of variables, creations of typologies are detailed in the data paper: Sandra Brée, Victor Gay, Marion Leturcq, Yoann Doignon, Baptiste Coulmont, et al. POPP. An OCR-Generated Database of the Population Censuses of Paris (1926-1936). Longitudinal and Life Course Studies, 2024. The variables followed by suffix '_c' are the corrected variables but the original, uncorrected variables are also provided. It is advisable to use the corrected variables. For some variables, part of the cells containing elements have not been corrected (e.g. when occurrences were less than 10) and a search can therefore be performed in the uncorrected variable. All these elements are detailed in the variable dictionary that accompanies the base and that it is necessary to read before using the base.

Variables

Not provided

Data

No data file is described in the record.

Documentation

Documents

No attached document in the record.

Publications related to the survey

Data access

SUS Unavailable

SUS : Cette référence n'est actuellement pas disponible. Pour plus d'informations, merci de nous écrire à diffusion.adisp@cnrs.fr

Number of files : 1 file

Request portal

  • Quetelet-Progedo data request portal

Access notes

  • The data are available in SAS, SPSS, STATA and CSV formats.

Contact

  • Adisp team

Archive originally storing the data

ADISP - Archives de Données Issues de la Statistique Publique

Confidentiality agreement

To obtain the data, the user must first submit a request using the Quetelet-Progedo-Diffusion data request portal; approval is subject to the user's status, as well as the presence of an institutional e-mail address and a research project. If a request is approved, an individualised agreement with ADISP (and in the case of FPR, a confidentiality agreement with the Statistical Confidentiality Committee) must be signed. Such agreements will require the user to adhere to the following conditions: -To use the data exclusively for research purposes; -Not to transfer the data, in any form whatsoever, to a third party, whether free of charge or for a fee; -To process these data in accordance with rules of art and statistical secrecy; -To mention the data source in their communications, publications, etc., according to the citation to be used whenever re-using the data (see below); -To inform the data distributor of any communications, publications, etc. they produce and to send the data distributor the references; -To inform the data distributor of any findings related to data quality or difficulties in re-use; -To inform the data distributor of any data re-use for research other than what was specified above; -To destroy the data files once the research has been completed; -To respect the regulations on the protection of personal data.

Restrictions

Files disseminated via Quetelet-Progedo-Diffusion are accessible to French and international researchers, doctoral students, post-doctoral students, and master's students for the purposes of research, scientific production and in some cases teaching. Any commercial use is prohibited. The term, "for the purposes of research," is defined as any scientific production or reproduction that validates new general knowledge. The results must be made public and free to disseminate.

Terms of use

The user must submit a request using the Quetelet-Progedo-Diffusion data request portal; approval is subject to the user's status, as well as the presence of an institutional e-mail address and a research project. If a request is approved, an individualised agreement with ADISP (and in the case of FPR, a confidentiality agreement with the Statistical Confidentiality Committee) must be signed. If the data are to be analysed by a team, all members of the research team must apply for access.

Required citation

doi:10.13144/lil-1719 Recensements de la Population Parisienne - 1926-1936 - CNRS (producteur), Brée, Sandra ; Paquet, Thierry ; Constum, Thomas ; Tranouez, Pierrick ; Gay, Victor ; Kempf, Nicolas ; Leturcq, Marion (auteurs), PROGEDO-ADISP (diffuseur).

Publication deposit

In order to provide the data producer(s) with essential information concerning how their data is being re-used, data users are required to notify ADISP (diffusion.adisp@cnrs.fr) of any publication produced as a result of exploiting the survey data, by sending, whenever possible electronically, the bibliographic reference and abstract.

Disclaimer

The data producer and data distributor cannot be held responsible for any interpretations or conclusions resulting from the re-use of the data.

Export the DDI

← Back to catalogue