skip to primary navigation skip to content

Constructing the first municipal-level GIS for France and a multi-modal transport network

Constructing the first municipal-level GIS for France and a multi-modal transport network


The COMMUNE HIS-DBD project will build the first historical-GIS capturing all changes in the boundaries of French communes since the Revolution and create a multi-modal dataset of transport networks from 1750 to the present.

It combines the strength and expertise of INED (team led by Isabelle Séguy), the Cambridge Group for the History of Population and Social Structure (CAMPOP) based at the University of Cambridge (team led by Max Satchell and Alexis Litvine), and the ThéMA lab at the Université de Bourgogne (team led by Thomas Thévenin). It also counts with the technical expertise of the Institut National de l'Information Géographique et Forestière (IGN) and the active support of some of the most prominent European scholars in our Advisory Board.

Previous attempts have not succeeded in producing a reliable and accurate database of boundary changes mostly because they did not rely on historical records to reconstruct historical administrative units. As nobody has yet undertaken the painstaking historical reconstruction on a national scale the exact shape and delimitation of French administrative units in the past remain largely unknown. The unique combination of the historical and cartographic expertise of our teams will allow us to fill this major historical gap. Thanks to the seminal work by Séguy and Théré, who compiled a list of all boundary changes since 1801, we have devised a method to reconstruct historical administrative boundaries efficiently and accurately. We will be using a combination of historical and cartographic sources drawing upon administrative records, official maps from both the Service Géographique de l'Armée (SGA) and IGN, and other cartographic material such as cadastral mapping, and recent land surveys. Overall c.15% of all communes, will require exhaustive archival research all over France in order to digitise contemporary maps of administrative boundaries.

Figures (1) to (4): Georeferencing the commune boundary between Louvignies-Bavay and Bavay

BD TOPO boundariesGeorectified map including corrections

(1) BD TOPO boundaries (2) Georectified map including corrections

Drawing historical boundariesResulting GIS boundary

(3) Drawing historical boundaries (4) Resulting GIS boundary

Over the years the Cambridge Group for the History of Population and Social Structure (CAMPOP) has developed significant expertise by reconstructing English and Welsh parish-level boundaries for the period 1831-1881 with sources scarcer than in the French case. The English and Welsh parish and place GIS is now the standard resource used by historians, geographers, archaeologists and economists for mapping historical data for England and Wales, and the GIS shapefiles that resulted from this work have been used by a great number of scholars in very different fields.

Thanks to a team of highly-qualified research associates and research assistants working both in France and in the UK we will put France on an equal footing with other European countries, and will unlock new fundamental comparative and quantitative analyses.

By the end of the research grant we will have produced the first and only set of GIS boundary data for all French communes from 1790 to the present linked to several key demographic, socio-economic and environmental data available at low level. Linking these data to fine spatial units through time and, critically, through them to each other will be a major breakthrough in the application of quantitative analysis of large datasets to French history. This will also enable us to compare spatial and historical variations in those key variables among several countries with a level of details and accuracy never reached before. Finally, our tool will be made available to the entire scientific community providing both transparency and traceability. Through a dynamic database linked to a set of annualised GIS shape files all historical records and maps used to reconstruct boundaries will be built in the data and accessible to all researchers.

Reconstructing historical transport networks in France, 1750 to the present

The existing geohistorical data on French transport infrastructures have been partially digitised by different institutions, with varying degrees of completeness: the railway network has been fully digitised from 1832 to 2015 with data on opening, closing dates and the evolution of technological innovation (electrification, rolling stock, high speed line…). The road network c.1750 has been digitised from Cassini maps (1750) by the COGIT laboratory at IGN (Perret et al., 2015). In this project we are: i) adapting existing railway data to fit to the much higher accuracy standards used for the GIS of French Communes; ii) digitising roads for two benchmark years (1830 and 1910). We use French military maps (Etat Major) for the period 1820 to 1866. For the second date, the 1900 maps (type 1900) digitised by IGN are our main source; iii) digitising navigable waterways 1700 to the present. The most consistent part of the data collection concerns French waterways. We use existing database of links, locks and tunnels, and complete it with data created by the Babel project. It includes construction dates, opening time and the technical evolutions for all existing canals. We complete this with data extracted from works by Delalande (1778), Grangez (1855) and Pinon (1886), allowing us to provide speed and cost assessment based on rights of way.

The French railway network since 1860 (Thévenin/Mimeur)

The French railway network since 1860 (Thévenin/Mimeur)

Building a multimodal model of transport accessibility in the long run

We are also creating homogeneous and comparable data on accessibility since 1750. This will greatly improve our understanding of the morphogenesis of transportation networks in France. In this way we will be able to quantify rigorously how changes in transportation networks contributed to spatial inequalities, and conversely we will be able to explore the extent to which public policy influenced network organisation.

The multimodal approach will allow us to combine microscopic and macroscopic analyses in order to look at the multi-scalar effects of spatial interactions. Combining roads, railways and waterways in the same model will help us assess the impact of innovations on territories suggested – but never quantified - by Dupuy (1991). We will provide measures of potential substitution effects and/or the complementarities of transportation networks. Finally, we will apply a morphological analysis of networks, in order to observe the impact of the co-evolution of transportation systems on population dynamics and notably emergence of centralities. These different hypotheses will be assessed by systematic comparisons between UK and France.

The effects of the railways on accessibility to Paris (Thevenin/Mimeur)

The effects of the railways on accessibility to Paris (Thévenin/Mimeur)

Description of the methodology

Read more


  • Isabelle Séguy, INED
  • Christine Théré, INED
  • Bénédicte Garnier, INED
  • Leigh Shaw-Taylor
  • Max Satchell
  • Alexis Litvine
  • Prof Thomas Thévenin, ThéMA - Université de Bourgogne
  • Dr Christophe Mimeur, ThéMA - Université de Bourgogne

Data use

With just over 6,000 boundary changes in two-hundred years reconstructing French historical boundaries is a manageable task, and it will lay the groundwork for many future historical enquiries. Since we launched our pilot project in January 2018, we have been contacted by numerous colleagues working in diverse fields inquiring about our data. In some cases, we were able to supply boundary data specially recreated for their projects:

- Schwartz (2010) used the French railway database to assess the effect of the railways on the restructuration of agriculture after the agrarian crisis.

- Henneberg (2013) used the railways GIS data for a comparative study of the role of the railways in the European integration over time.

- Dehdari and Gehring (2018) use the new information about administrative changes to evaluate the credibility of exploiting a natural historical experiment in the French-German border regions. This was previously impossible, but geo-referencing the changes and linking them to historical census and other data enables them to conduct meaningful estimates at the micro-level. More specifically, they use information on population and population density to evaluate whether the drawing of a border following the Franco-Prussian war was related to demographic factors.

- Heblich, Redding and Sturm (2019) use the commune population data to derive weights for a consistent definition of the center of Paris before and after the Haussmann reforms in 1860. The rearrangement of arrondissements led to substantial area changes which make it hard to justify area weights to derive a consistent population measure. Breaking population down to Quartiers and Communes is a far more credible way to generate consistent spatial units over time. We use this information to test the external validity of our findings for London which suggest that the rise of transportation technologies that facilitated commuting led downtown areas to depopulate while surrounding areas in commuting distance increase in population.

- Ryavec (2019) has used our data to develop a case study between France and China for a geohistorical course (Mapping political history over time).

Other endorsements:

- Prof Eric Brian: 'Ce projet servira à tou(te)s les historien(ne)s qui voudront examiner les phénomènes démographiques et économiques de l'époque moderne en étant dès lors en mesure d'en saisir de manière réaliste la dimension spatiale. En son principe, il servira sans doute d'exemple pour d'autres pays ... C'est pourquoi j'appuie sans réserve ce projet, et s'il venait à aboutir, je peux annoncer dès aujourd'hui que je m'y associerais avec enthousiasme : pour ma part ce serait une nouvelle étape dans une recherché conduite depuis 1990, soit près de trente années, et dans la perspective plus longue qu'offre l'histoire des sciences, ce serait l'aboutissement de travaux commencés à Paris vers 1770, soit il y a près de 250 ans !'

If you would like to use our boundary data or request specific boundary data for one of your projects please contact Alexis Litvine (


  • Généanet
  • Université de Bourgogne
  • INED

Funding is being sought as part of project COMMUNE HIS-DBD (COllaborative Micro Mapping of UNExploited HIStorical District-Boundary Data), submitted to the Agence Nationale de la Recherche (ANR, France).

INED logoTHEMA logoGeneanet logo