Observation of the territory for the understanding of the evolution of administrative boundaries and settlements. Building Geo-Historical data bases and Ontological approach. – GéoPeuple

History of the French population and territory since the 18th century until now is quite well known at the department's and province's level. However, it is far less explored at the scale of the smallest administrative units, the communes, where the people have their everyday life. The reason for that lies in the bulk of data to be handled: for twenty or so provinces and regions or for a hundred departments, they are more than 40 000 communes appearing during the last two centuries and a similar number of parishes before the Revolution. The number of people in each commune has to be collected in the censuses; administrative borders of the communes which no more exist has to be drawn; places of economic, military and demographic interest indicated on the ancient maps have to be taken into account, geo-referenced and followed through the time. Only a precise knowledge of these elements can help understand how administrative units, local population growth and buildings of interest interact and as a consequence, help anticipate consequences of changes of the administrative grid as actually foreseen.
The GeoPeuple project is intended to afford, from one hand, a technical tool to gather geo-referenced data at the local level using ancient maps an censuses, and on the other hand, to perform an historical analysis of the population at the local level in relation to the administrative structure and the characteristic elements of the physical and human landscape (roads, forests, marshes, factories, mills, fortifications...). The project is feasible because the three teams involved join their ability and their experience in their own field. The historical demography laboratory (LDH) at the EHESS has gathered demographic data since 1791 at the commune's level and is finishing estimating the border of disappeared communes. It has also rebuilt in raster mode the Cassini maps going back to the mid 18th century with the help of the Bibliothèque Nationale. The LDH will provide this raw material together with an interpretation of the results. The LIP6 of Paris 6 University is well known in IA techniques and image processing. Its task will be to build algorithms and software able to recognize and vectorise special items on the maps. These elaborated tools must have a degree of generality concerning the kind of maps so as to be used by the COGIT of the IGN specialized in computer geography which will follow the selected items from old map to current geo data base (RGE), elaborating rules of division of the territory and rules of change and also a system of geo-identifier taking into account the time. In this way, the COGIT will be at the center of the project, linking the software and algorithms elaborated by the LIP6, to the historical interpretation of the LDH. In this respect, the definition of the rules governing border, places, population, and their test using simulations will be strategic. Together with the LDH, the COGIT will popularize the results in an easily understandable way pointing in three directions. First, the government interested in knowing the conditions of stability of a new administrative grid will take advantage of the analysis of more than 10 000 communes’ fusions and fissions which occurred in the past. Second, the historians, geographers and demographers working on the territory will find data and tools aiming at identifying and gathering any symbol on the maps and relating it by georeference to the whole administrative grid as well as to local populations and other items present at a given time. Third, the common people will find some answers on how the french population occupied its territory at a fine level since more than two centuries and also a general view of how the major changes of population and administrative centers took place during the last 220 years.