Improving geospatial data usability

点击翻译按钮获取中文摘要

Geographic Information Systems (GIS) are mapping software increasingly being used to support decision-making in various fields, ranging from social and human sciences (e.g. anthropology, sociology) to engineering (e.g. civil engineering). Indeed, according to certain studies, more than 80% of data can be spatially referenced, and taking the geographic environment into account may be critical in many cases to better understand the phenomenon analysed. However, the data used by the GIS are always, at different levels, inaccurate, imprecise, incomplete or out-of-date, being one representation of the geographic environment at a given time, produced using acquisition procedures of various qualities (e.g. land surveying, GPS, satellite imagery). We would then expect GIS to take into consideration these uncertainties in their calculations, but they unfortunately process data as if they were perfect, even if quality is often well-known. Information about quality is often produced by data producers and is more and more frequently provided together with data as part of the overall metadata (data about data). If most end-users consider GIS results as true, expert users know there is some level of uncertainty, but usually can't estimate this level. It is now common to observe problems that result from a decision based on GIS using geospatial data of low quality. Such decisions can have strong economic, social or political impacts and already led to the death of people. With the increasing number of users, it is then critical to be able to provide users with information about what they can expect from geospatial data. The aim of this research program is to enhance the usability of geospatial data. This will be done by developing techniques to improve the communication of quality information to the end-users as well as designing tools that would increase the use of data quality information by GIS in order to reduce the risks of misuse of geospatial data by the end-users.