Geoffrey Brun

At COGIT team from 2011 to 2015

Phone : 33 1 43 98 80 00 + 72 19

-

Research topic in the laboratory COGIT :

With a view to design maps according to user’s needs, this PhD work assume to find these needs using a text provided by this user. Map design from a text might be guided by several elements, like thematic information and location names occurring in the text. These elements are clues to create map key, map extent, map administrative units, map title or even map style.

In this PhD work, two elements are studied: the map extent and the mapped administrative units. The first one is the maximum geographic coordinates “north/south/west/east” of the map. The second one is constituted by the different mapped administrative levels (continents, countries, counties, cities, etc.). Text clues which might guide automatic creation of these two map elements are mainly locations name occurring in this text. These last could be more or less easy to identify, e.g. Madrid, Southern China, the east of Russia, the city of Paris, the Kuril Archipelago or even the western edge of the Eastern Siberian plateau. Location names ambiguity is also an important problem to take into account: for example, Paris could refer to several cities (in France and in the United-States) and Niger could be a country or a river.

The proposed method consists first in extracting location names occurring in a text in the form of Spatial Named Entities (SNE). These SNE, which are only strings, are then matched with toponyms, holding semantic information, stored in a gazetteer by using disambiguation algorithms. Afterwards, these toponyms are matched with geographic entities (points, lines and polygons) stored in a database. Various algorithms are finally applied on this list of geographic entities in order to create a map extent and define administrative units to map (count and grouping of the entities based on semantic criteria, spatial relations, etc.).

Education :

On-going PhD in geography (Université Paris Diderot - Paris 7):
From text to map: create the spatial extent and administrative units of a map from location names occurring in a text
Supervised by Christine Zanin, co-supervision : Catherine Dominguès

Professional experience :

  • Mars-September 2011: professional training in geomatics at the Haut-Languedoc Regional Natural Park. Needs assessment in geographic information and development of a web cartographic application (using PostGreSQL/GeoServer/GeoWebCache/GeoExt).
  • February-May 2010: professional training in town and country planning at the Central European Institute of Technology (CEIT ALANOVA) in Schwechat (Austria). Involvement in the ICT-ENSURE project (study on ICT applied to town and country planning).

Skills :

  • Cartography & semiology.
  • GIS software & spatial databases.
  • Web mapping: cartographic servers & clients, web services WMS & WFS.
  • Natural Language Processing (Unitex, TreeTagger).
  • Textual statistics (iRaMuTeq, RTmétis).
  • Algorithmic & development (HTML, CSS, PHP, JavaScript, JAVA).
  • Languages: English, German.

Publications :

- All publications of Geoffrey Brun

Teaching given at COGIT :

  • ,
    M2 DDMEG, ENSG. 2013.
  • Webmapping projects: OpenLayers, PostGIS, GeoServer,
    Master Géomatique/Carthagéo, ENSG. From 2012 to 2014.
  • Cartography and semiology,
    Formation continue, ENSG. From 2012 to 2013.
  • QGIS,
    Master Géomatique/Carthagéo, ENSG. 2012.
  • Introduction to GIS, MapInfo and ArcGIS,
    M2 Géorisques, Université Marne-la-Vallée. 2012.
  • Intoduction to GIS and ArcGIS,
    M2SIAW, Université Marne-la-Vallée. 2012.
--
-
IGN Web Site